Protection Control

PC2_cover 10/21/04 9:43 AM Page 1
Buyer: Esam Darweesh (esamdrsh@gmail.com)

Transaction ID: 43E17990X7838060L
Electrical System Protection and Control Handbook Volume 2

MTE Introduces the Full Spectrum
In motor protection:
Sine Wave Filters and dV/dt Filters
New MTE Sine Wave filters provide
sine wave output voltage when driven
from PWM inverters, eliminating
motor insulation failures, reduce EMI
and help meet IEEE-519 for cable
lengths up to 15,000 feet
New MTE Series A sine-wave filters provide
a sine-wave output voltage when driven from
PWM inverters with switching frequencies from
2 kHz to 8 kHz. For drive applications, these
filters eliminate the problem of motor insulation
failures and reduce electromagnetic interference
by eliminating the high dV/dt associated with
inverter output waveforms in applications
where the distance between the motor and
the inverter is up to 15,000 feet. MTE Sine
Wave Filters permit the use of standard
MTE Sine Wave Filters transformers and meet the requirements of
IEEE—519.
First-turn arc-over protection GUARANTEED for
cable lengths up to 1,000 feet between inverter
and motor with new MTE dV/dt filters
MTE Series A dV/dt filters are designed to protect AC motors
from the destructive effects of peak voltages facilitated by
long cable runs between the inverter and motor. The MTE
dV/dt filter is guaranteed to meet its maximum peak motor
voltage specification (150% of bus voltage) with up to 1,000
Volume 2
feet of cable between the filter and the motor. It is also rated
for a maximum dV/dt of 200 volts per microsecond. In specific
applications, the filter has provided excellent performance
with cable runs up to 3,000 feet. The dV/dt filter has a 3%
insertion impedance which ensures motor torque is not The Electricity Forum
affected by added voltage drops from the filter.
MTE dV/dt Filters
Visit us at: www.mtecorp.com

Or call 1-800-455-4MTE (4683)
THE INTERNATIONAL POWER QUALITY EXPERTS

PC2_cover 10/21/04 9:43 AM Page 2

CONNECTING
...PROTECTING
®
®
From overhead to underground...
Hubbell has the products and know-how you need for
planning, constructing and maintaining distribution and
transmission networks. With traditional and innovative
products, Hubbell helps you cut costs, save time and
increase efficiency.
For transmission and distribution, look no further than

Hubbell. We have the products, support and expertise to
ensure your competitive advantage.
Along with the many advantages of
deadfront, Chance AIS Padmounted
Air-Insulated Switchgear also offers a
visible break, fusing and switch configuration
choices, motorized operation, and reduced
long-term maintenance costs.
POWER SYSTEMS 870 Brock Road South • Pickering, ON L1W 1Z8 • Phone (905) 839-1138 • Fax: (905) 831-6353 • www.HubbellPowerSystems.ca
pc_v2 10/19/04 10:24 AM Page 1

Electrical System
Protection and Control
Handbook
Volume 2
Published by The Electricity Forum
The Electricity Forum The Electricity Forum Inc.

204 -15 Harwood Avenue South One Franklin Square, 4th Floor
Ajax, Ontario L1S 2B9 Geneva, New York 14456
Tel: (905) 686-1040 Fax: (905) 686 1078 Tel: (315) 789-8323 Fax: (315) 789 8940
E-mail: hq@electricityforum.com E-mail: forum@capital.net
Visit our website at

w w w. e l e c t r i c i t y f o r u m . c o m
pc_v2 10/19/04 10:24 AM Page 2

2 Electrical System Protection and Control Handbook - Vol.2
Electrical System
Protection and Control
Handbook
Randolph W. Hurst
Publisher & Executive Editor
Phill Feltham
Editor
Art Design
Alla Krutous
Handbook Sales
Colleen Flaherty
Elena Hurst
Advertising Sales
Anita Faiella
Carol Gardner
Barbara John
Database Administration
Hendra Sianto
The Electricity Forum

A Division of the Hurst Communications Group Inc.
All rights reserved. No part of this book may be reproduced without
the written permission of the publisher.
printed and bound in Canada.
ISBN-0-9735726-3-9
The Electricity Forum
204 - 15 Harwood Avenue South, Ajax, ON L1S 2B9
Printed in Canada
© The Electricity Forum 2004

pc_v2 10/19/04 10:24 AM Page 3

Electrical System Protection and Control Handbook - Vol.2 3
TABLE OF CONTENTS
Modern Cost-Efficient Digital Busbar Protection Solutions
By Bogdan Kasztenny and Gustavo Brunello, GE Multilin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5
Testing Consideration for Modern Distance Relays

By Bogdan Kasztenny, GE Multilin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19
Vector Jump Relaying: An Economical Solution for Distributed Generation Islanding Protection
By Arvind Chaudhary, Cooper Power Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27
Asset Management: Purchasing Data Integration for Relay Protection

By Joseph Stevenson, ENOSERV LLC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31
Protection and Control Upgrade -User Added Value

By Schneider Electric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35
Analyze Relay Fault Data to Improve Service Reliability

By Roy Moxley, Schweitzer Engineering Laboratories, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37
How Certain Features of Protective Devices Help to Protect Workers

By Jeanne K. Ziobro, Eaton Corporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .45
Choosing the Right Solid State Relay

By Lyle W. Strode, HBControls, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .47
Aspects of Overcurrent Protection for Feeders and Motors

By E. 0. Scheitzer III and S. E. Zocholl, Schweitzer Engineering Laboratories, Inc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .51
Thermally Protected MOV

By Ferraz Shawmut . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57
Rio Grande Electric Monitors Remote Energy Assets Over Satellite

By Anthony Tisot, Power Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .61
Generator Protection Application Guide

By George Rockefeller, Basler Electric Company . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63
Integrated Protection, Metering and Control used with Monitoring using web enabled communication technologies
By Ajit Bapat, M.Tech, MBA, P Eng, Schneider Electric Canada . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .79
A new approach to high impedance fault detection

By Ratan Das, Ph.D., P.Eng, ABB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .83
A Look at Fuseology
By Tim Crnko, Cooper Bussman . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .89
BUYERS GUIDE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .93

pc_v2 10/19/04 10:24 AM Page 4


pc_v2 10/19/04 10:24 AM Page 5

MODERN COST-EFFICIENT
DIGITAL BUSBAR PROTECTION SOLUTIONS
By Bogdan Kasztenny and Gustavo Brunello, GE Multilin
INTRODUCTION (d) Several differential zones are required to cover individual
Simple busbars with dedicated Current Transformers sections of a large bus. This calls for significant processing
(CTs) could be efficiently protected by the high-impedance power of the hardware platform.
principle – a fast and reliable scheme with tens of years of Traditionally, the aforementioned problems of large num-
excellent field record. However, new power generation added ber of inputs and outputs, resulting power supply requirements,
recently, or to be added in the near future, complicates histori- and the processing power requirements have been addressed by
cally simple busbar arrangements and exposes existing CTs to two distinctively different architectures: distributed or central-
saturation due to increased fault current level. New substations ized. Both solutions require large Modern Cost-Efficient Digital
are often designed to satisfy cost requirements rather than keep Busbar Protection Solutions quantities of specialized hardware.
the protection task straightforward and easy. This results in As a result, they are difficult to engineer, face certain depend-
complex busbar arrangements. ability and reliability problems, do not have a chance to mature
High-impedance busbar protection principle faces major due to comparatively low volume of installations, and are very
problems when applied to complex busbar arrangements. Quite expensive.
often, the zones of protection are required to adjust their bound- The new solutions that emerged recently address the
aries based on changing busbar configuration. This calls for above problems by targeting medium-sized busbars only and
switching secondary currents – an operation that is never con- using generic hardware platforms (such as multi-winding or
sidered safe and should be avoided whenever possible. small bus relays) to build phase-segregated, cost-efficient, easy-
Digital low-impedance busbar protection schemes are to-engineer busbar relays.
ideal for complex busbars. Optimal zoning (dynamic bus repli- This paper focuses on the new phase-segregated solution
ca) is achieved naturally by switching currents in software, i.e. and is organized as follows.
by making logical assignments to multiple zones of protection First, a general overview of busbar protection principles
while keeping physical currents uninterrupted. Other benefits is given starting from simple interlocking schemes for single-
include integrated breaker fail protection, communications, incomer distribution busbars, to high-end microprocessor-based
oscillography, sequence of events recording, multiple setting protection schemes.
groups, and other natural benefits of the digital generation of Second, a novel, phase-segregated approach based on
protective relays. existing hardware platforms capable of processing plurality of
Until very recently, digital busbar and breaker failure single-phase AC input signals is presented. The new solution is
protection schemes for medium-size and large busbars were not discussed in detail including architecture, reliability, depend-
attractive to users traditionally biased toward the high-imped- ability, speed of operation, security on external faults, ease of
ance approach. There used to be several reasons for that. configuration, and cost.
Schemes available on the market were very expensive, difficult Third, basic application principles for protection of com-
to apply, considerably slower as compared with the high-imped- plex busbars are presented. They include a tie-breaker with a
ance protection, and perceived less secure. All these factors single CT, treatment of blind zones and over-tripping zones,
have changed recently. Modern digital relays are much faster, dynamic bus replica, end fault protection and breaker failure
use better algorithms for security, and became affordable after protection. Both principles and examples are presented.
introduction – in late 2001 and early 2002 – of a phase-segre-
gated microprocessor-based busbar relay. BUSBAR PROTECTION TECHNIQUES
Major hardware, architectural and processing power Power system busbars vary significantly as to their size
challenges facing a digital protection system for medium-size (number of circuits connected), complexity (number of sections,
and large busbars are: tie-breakers, isolator switches/disconnectors, etc.) and voltage
(a) Large number of analog signals needs to be processed (tens level (transmission, distribution).
of currents, few voltages). The problem is how to bring all The above technical aspects, combined with economic
the required signals into a “box”. factors, yield a number of protection solutions.
(b) Large number of digital inputs may be required to monitor
isolator and breaker positions in order to provide for the INTERLOCKING SCHEMES
dynamic bus replica mechanism (dynamic adjustment of A simple protection for distribution busbars can be engi-
zone boundaries based on changing busbar configuration). neered as an interlocking scheme. OverCurrent (OC) relays are
(c) Large number of trip-rated output contacts may be required placed on an incoming circuit and at all outgoing feeders. The
particularly in the case of reconfigurable busbars when feeder OCs are set to detect feeder faults. The OC on the incom-
each breaker must be tripped separately depending on bus ing circuit is set to trip the busbar unless blocked by any of the
configuration at the moment of tripping. feeder OC relays (Figure 1). A short coordination timer is
pc_v2 10/19/04 10:24 AM Page 6


Although economical and applicable to distribution bus-
bars, this solution does not match performance of more
advanced schemes and should not be applied to transmission-
level busbars.
The principle, however, may be available as a protection
function in an integrated microprocessor-based busbar relay. If
this is the case, such unrestrained differential element should be
set above the maximum spurious differential current and may
give a chance to speed up operation during heavy internal faults.
PERCENT DIFFERENTIAL
Percent differential relays create a restraining signal in
addition to the differential signal and apply a percent
(restrained) characteristic. The choices of the restraining signal
include “sum”, “average” and “maximum” of the bus currents.
The choices of the characteristic typically include single-slope
and double-slope characteristics.
This low-impedance approach does not require dedicated
Fig.1. Illustration of the interlocking scheme. CTs, can tolerate substantial CT saturation and provides for
comparatively high-speed tripping.
required to avoid race conditions. Many integrated relays perform CT ratio compensation
When using microprocessor-based multi-function relays eliminating the need for matching CTs.
it becomes possible to integrate all the required OC functions in This principle became attractive with the advent of
one or few relays. This allows not only to reduce the wiring but microprocessor-based relays because of the following:
also to shorten the coordination time and speed up operation of • Advanced algorithms supplement the percent differential
the scheme. protection function, making the relay very secure.
Modern relays provide for fast peer-to-peer communica- • Protection of re-configurable busbars becomes easier as the
tions using protocols such as the UCA with the GOOSE mech- dynamic bus replica (bus image) can be accomplished with-
anism. This allows eliminating wiring and sending the blocking out switching secondary currents.
signals over digital communications. • Integrated Breaker Fail (BF) function can provide for opti-
The scheme, although easy to apply and economical, is mum tripping strategy depending on the actual configura-
limited to simple distribution busbars. tion of a busbar.
• Distributed architectures could be used that place Data
OVERCURRENT DIFFERENTIAL Acquisition Units (DAU) in bays and replace current wires
by fiber optic communications.
Typically, a differential current is created externally by
summation of all the circuit currents and supplied to an overcur-
rent relay (Figure 2). Preferably the CTs should be of the same
HIGH-IMPEDANCE PROTECTION
ratio. If not, matching CTs are required. This in turn may High-impedance protection responds to a voltage across
increase the burden for the main CTs and make the saturation the differential junction points. The CTs are required to be of a
problem even more significant. low leakage (completely distributed windings or toroidal coils).
Historically, means to deal with the issue of CT satura- During external faults, even with severe saturation of some of
tion include definite time or inversetime overcurrent character- the CTs, the voltage does not rise above a certain level, because
istics. the other CTs will provide a lower-impedance path as compared
with the relay input impedance. This principle has been used for
more than half a century because it is robust, secure and fast.
The technique, however, is not free from disadvantages.
The most important ones are:
• The high-impedance approach requires dedicated CTs (sig-
nificant cost associated).
• It cannot be easily applied to re-configurable busbars
(switching currents with bi-stable auxiliary relays endan-
gers the CTs, jeopardizes security and adds an extra cost).
• The scheme requires only a simple voltage level sensor. If
BF, event recording, oscillography, communications, and
other benefits of microprocessor-based relaying are of
interest, then extra equipment is needed (such as a Digital
Fault Recorder or dedicated BF relays).
PROTECTION USING LINEAR COUPLERS

A linear coupler (air core mutual reactor) produces its
output voltage proportional to the derivative of the input cur-
Fig.2. Unrestrained differential protection. rent. Because they are using air cores, linear couplers do not sat-
pc_v2 10/19/04 10:24 AM Page 7


would recognize CT saturation on external faults in a fast and
reliable manner.
(c) Applying a second protection principle such as phase
directional (phase comparison) for better security.
Digital relays for large busbars dominating the market
until recently provide for a trip time in the range of 0.75 to 1.5
power cycles, and use either phase comparison principle or
decaying restraining current for increased security on external
faults. They were designed several years ago based on technol-
ogy that is now outdated by several generations of microproces-
sors.
In the meantime several new busbar relays have been
introduced based on modern, more powerful hardware plat-
forms [7,8]. These relays provide for faster tripping time and
modern features, but till recently their capabilities were limited
to small (typically six-circuit) busbars.
This changed with the introduction of digital, phase-seg-
Fig. 3. Busbar protection with linear couplers. regated, low-impedance protection schemes.
urate.
During internal faults, the sum of the busbar currents, and
thus their derivatives, is zero. Based on that, a simple busbar
protection is achieved by connecting the secondary windings of
the linear couplers in series (in order to respond to the sum of
the primary currents) and putting a voltage sensor to close the
loop (Figure 3).
Disadvantages of this approach are similar to those of the
high-impedance scheme.
MICROPROCESSOR-BASED RELAYS
The low-impedance approach used to be perceived as
less secure when compared with the high-impedance protection.
This is no longer true as microprocessor-based relays apply
sophisticated algorithms to match the performance of the high-
impedance schemes [1-6]. This is particularly relevant for large,
extra high voltage busbars (cost of extra CTs) and complex bus-
bars (dynamic bus replica) that cannot be handled well by high-
impedance schemes.
Digital low-impedance relays could be developed in one
of the two distinctive architectures:
• Distributed busbar protection uses DAUs installed in each Fig. 4. Distributed busbar protection.
bay to sample and pre-process the signals and provide trip
rated output contacts (Figure 4). It uses a separate Central
Unit (CU) for gathering and processing all the information
and fiber-optic communications between the CU and DAUs
to deliver the data. Sampling synchronization and/or time-
stamping mechanisms are required. This solution brings
advantages of reduced wiring at the price of more complex,
thus less reliable, architecture.
• Centralized busbar protection requires wiring all the sig-
nals to a central location, where a single “relay” performs
all the functions (Figure 5). The wiring cannot be reduced
and the calculations cannot be distributed between plurali-
ty of DAUs imposing more computational demand for the
central unit. On the other hand, this architecture is per-
ceived more reliable and suits better retrofit applications.
Algorithms for low-impedance relays are aimed at [8]:

(a) Improving the main differential algorithm by provid-
ing better filtering, faster response, better restraining technique,
robust switch-off transient blocking, etc.
(b) Incorporating a saturation detection mechanism that Fig. 5. Centralized busbar protection.
pc_v2 10/19/04 10:24 AM Page 8

PHASE-SEGREGATED BUSBAR RELAYS lower as compared with traditional, “specialized” digital busbar
The problem of a large number of inputs and outputs relays. Other features, benefits and peculiarities of the phase-
required for protection of medium-size and large busbars, as segregated approach are discussed in subsection 3.6.
well as computational power required to perform all the neces- Busbar protection is more than a plain differential func-
sary operations on the inputs, could be solved by using a phase tion. The following subsections address several issues related to
segregated approach to busbar protection. features such as breaker failure protection, under-voltage super-
From the perspective of the main differential protection, vision, dynamic bus replica, etc.
the algorithm is naturally phase-segregated. This means that no
information is required regarding currents in phases B and C in DIFFERENTIAL PROTECTION
order to fully protect phase A. This bears several important con- The main differential protection function is implemented
sequences and advantages. on a per-phase basis. A given solution could be more flexible,
First, completely independent microprocessor-based cost-efficient, and allow more demanding applications, if multi-
devices could process the AC signals that belong to phases A, B ple zones of protection are available.
and C. No data transfer is required between the devices. A full-featured digital busbar protection system incorpo-
Second, sampling synchronization is not required rates dynamic bus replica function. This includes both ability to
between the separate devices processing signals that belong to dynamically assign currents to the relay differential zones, and
individual phases. provide for reliable monitoring of the status signals for isolator
The above observations facilitate phase-segregated bus- switches and breakers. The latter is typically implemented by
bar protection. With reference to Figure 6, three separate relays utilizing both normally open and normally closed auxiliary
(Intelligent Electronic Devices, IEDs) could be used to set up switches as explained in subsection 3.4.
protection for a three-phase busbar. Each device is fed with AC Under-voltage supervision (release) of the main differen-
signals belonging to the same phase, processes these signals, tial function is an often-used feature. This feature guards the
and arrives at the trip/no-trip decision. On solidly grounded sys- system against CT trouble conditions and problems with the
tems, at least one device would operate for any type of fault. For dynamic bus replica (false position of a switch/breaker).
phase-to-phase faults, two relays would operate. Typically, phase under-voltage, or neutral and/or negative-
In order to protect a medium-sized busbar, it is enough sequence over-voltage functions are used. The phase-segregated
that each IED supports some 18-24 AC inputs. Present protec- approach treats single-phase voltage inputs in a generic way.
tion platforms support this amount of AC inputs. A modern The user could wire phase voltages or neutral (broken delta VT)
multi-winding transformer, or small busbar relay, could thus be voltages for the purpose of voltage supervision to all, or select-
converted from a three-phase device into a single-phase differ- ed IEDs only. As a rule, a voltage abnormality in any phase
ential device. Instead of supporting four or more three-phase releases all three phases of differential protection. This calls for
inputs, and some ground inputs, the relay supports 18-24 gener- simple inter-IED communications. This could be done via input
ic AC inputs and allows for configuring them as inputs to the contacts, or digital inter-IED communications means.
same zone of differential protection. In the phase-segregated approach each phase IED could
This approach yields a number of significant benefits. drive an output contact for the trip command. This could be
The three most important ones are: First, by building on exist- done on a per-breaker basis, if required. External lockout relays
ing platforms, the vendors could develop such a solution in a may be used to gather the per-phase trip commands in order to
short time with a very low investment. Second, utilizing stan- generate a single three-pole trip signal, if required.
dard platforms brings extra maturity and features into the bus-
bar applications. Third, by building on standard hardware plat- INTER-RELAY DIGITAL COMMUNICATIONS
forms, the manufacturing cost is also reduced. Consequently, Eliminating the AC data traffic between the devices facil-
the overall cost of the phase-segregated solution is substantially itates the digital phase-segregated busbar protection scheme. It
is very beneficial, however, to provide for fast, reliable, fully
Fig. 6. Phase-segregated digital busbar scheme.

pc_v2 10/19/04 10:24 AM Page 9


programmable communications mean for sharing on/off states BREAKER FAILURE PROTECTION
between the IEDs comprising the busbar protection system.
Integrated Breaker Failure (BF) protection is quite bene-
Important applications of such a communications mean are as
ficial for complex busbars. The BF system must monitor config-
follows:
uration of a busbar at any given time, in order to initiate appro-
• A large number of I/O points could be freely distributed
priate trip action upon a failure of a given breaker. Basically,
between the devices. This means that the same information
this information is identical with the information required by the
that is common for all three-phases, such as position of a
main differential protection. Doubling this information, i.e.
motorized switch, could be wired just to one IED, and dis-
using two separate systems for differential protection and break-
tributed to the remaining IEDs via digital communications.
er failure, is not economical, as a very significant number of I/O
• Applications such as cross triggering of oscillography,
points is required to interface the scheme with auxiliary switch-
under-voltage supervision, checkzone could be easily
es of all the isolator switches and breakers. This leads towards
accomplished by sharing on/off signals via communica-
integration of the busbar and breaker failure protection func-
tions channels.
tions.
• Certain breaker failure architectures rely heavily on digital
The BF imposes a design challenge for phase-segregated
communications.
approaches. Large number of currents must be monitored, large
• Certain dynamic bus replica solutions rely on communica-
number of I/O points must be monitored or driven, and large
tions as well. Tens or hundreds of digital inputs required for
computational power is required to process those signals.
monitoring all the switches and breakers, could be wired to
One particular solution monitors the currents using the
an extra relay (relays), where basic filtering (contact dis-
phase IEDs and sending the corresponding pickup/dropout sig-
crepancy and alarming) is performed, and the final status
nal via digital communications. This means that the devices
signals for the differential protection is sent via communi-
wired to a given signal are responsible for monitoring the level
cations.
of the signal and informing the BF-dedicated IED regarding the
One particular solution [9] uses redundant token-ring-
level of the current.
like communications mechanism as shown in Figure 7. Up to
With reference to Figure 8, devices 1, 2 and 3 are main
eight devices could be connected. Each device could share up to
protection devices for phases A, B and C, respectively; while
96 points with all the other devices. The mechanism is based on
device 4 is a BF-dedicated device. Devices 1, 2, and 3 constant-
“auto-forwarding”: each message is received, decoded if
ly feed the fourth device with information regarding the current
required, and forwarded to the next device. When the originat-
level. This does not create any extra bandwidth requirements as
ing device receives its own message back, it stops forwarding it.
only changes of the states initiate transmission. Device 4 hosts
This is also a sign that the communications ring is healthy. Two
all the BF schemes for all the breakers. Typically, this device is
rings increase reliability and reduce (by half) the maximum
configured to interface all the required I/O points. The BF func-
message delivery time. Available physical media include fiber,
tions could be initiated externally (typically via contact input),
RS422 and G.703. As the devices are connected directly, the
or internally (typically via communications). If any of the BF
820 nm fiber connection is typically used. 32-bit CRC is imple-
functions times out and operates, it closes an output contact on
mented for security, and the messages are repeated for extra
device 4, or sends a signal via communications to close any ded-
dependability. The maximum delivery time between two neigh-
icated contact located on any of the remaining devices of the
boring devices is 2-3 msec. Broken-ring alarm is incorporated
scheme.
for overall reliability of the scheme. Default states are available
For simple busbars, external BF relays could be used.
in order to program response of the scheme, should the commu-
When the busbar is re-configurable, however, it is always bene-
nications fail.
ficial to have the BF function integrated with the dynamic bus
Fig.7. Inter-IED communications for sharing on/off signals.

pc_v2 10/19/04 10:24 AM Page 10

Fig. 8. Inter-IED communications for sharing on/off signals.
replica function of the main differential protection. cally) or a failure to trip (rarely). Therefore, an isolator monitor-
ing element should respond to both normally open and normal-
DYNAMIC BUS REPLICA ly closed auxiliary contacts of an isolator or a tie-breaker in
Dynamic bus replica feature is critical for re-config- order to assert the actual position of the isolator for the dynam-
urable and complex busbars. An actual bus image must be monic bus image. Ideally, the element should assert two extra out-
itored by the protection system for the following purposes. puts for isolator alarm (contact discrepancy), and for blocking
First, for circuits with a single CT point that could be switching operations in the substation.
routed between different sections of the busbar, position of the Traditionally, the following logic is applied.
isolator switch must be known in order to dynamically decide if TABLE 1. STANDARD ISOLATOR MONITORING LOGIC..
the CT current belongs to a given zone of protection.
Isolator Open Isolator Closed Block Switching
Second, for circuits with a single CB point that could be Isolator Position Alarm
Auxiliary Contact Auxiliary Contact Operations
routed between different sections of the busbar, position of the Off On CLOSED No No
isolator switch must be known in order to dynamically decide if
Off Off LAST VALID After time delay Until Isolator
the CB should be tripped upon operation of a given zone of pro- until acknowledged Position is valid
On On CLOSED
tection.
On Off OPEN No No
Third, the BF protection should monitor all the breakers
connected to a given breaker, in order to decide a tripping strat-
egy should the said breaker fail. Typically, an alarm is set when contact discrepancy is
Fourth, positions of breakers and tie-breakers should be detected. Depending on the type of discrepancy, either the last
monitored in order to avoid blind spots or over-tripping zones. valid isolator position is assumed, or a “close” position is
Fifth, certain complex switching strategies call for signif- declared.
icant re-adjustments of zone boundaries. The re-adjustment It may be beneficial to block switching operations in the
should be programmed as a response to the changing configura- substation should a problem with the bus image occur. An oper-
tion of the busbar. ator could remove such blocking signal once the nature of the
Section 4 addresses the aforementioned application con- problem is discovered and rationalized.
siderations. Here, the aspect of dynamic zone boundaries and
the isolator monitoring feature are discussed. MODULARITY
Dynamic zone boundary could be programmed using a Some phase-segregated solutions offer extra modularity
very straightforward mechanism of configuring a zone of pro- at the IED level [9]. This may include variable number of DSPs,
tection as a list of pairs: current input – on/off status signal. A and I/O cards as well as communications and redundant power
given current becomes a part of the zone only if the correspon- supply. This brings an advantage of shaping each IED of the
ding status signal is asserted. Such a mechanism allows for busbar protection system to fit the needs of a particular applica-
maximum flexibility as the connection status signals could be tion.
freely programmed in user-programmable logic of the relay With reference to Figure 9b, an IED may be configured
based on a number of conditions, and different protection with 2 DSPs only, allowing measuring 2 x 8 currents, thus pro-
philosophies. tecting busbars of up to 16 breakers. Figure 9c presents a sam-
Auxiliary switches could fail to respond correctly. This is ple configuration with 3 DSPs (24 AC inputs), 3 I/O modules
particularly true for motorized switches. Wrong assignment of a (up to 3 x 16 inputs, or 3 x 8 outputs) and digital communica-
given current to a given zone – caused by incorrect information tions card. Figure 9d shows a configuration aimed at interfacing
regarding bus configuration – could result in a false trip (typi- exclusively I/O points, with no DSPs, but 5 I/O modules, com-
pc_v2 10/19/04 10:24 AM Page 11

Fig. 9. Sample configurations of a modular system: (a) power supply and CPU are manda- Fig. 10. Three-phase protection for small busbars.
tory, (b) 2 DSP, 2 I/O configuration for up to 16 current inputs, (c) 3 DSP, 3 I/O configura-
tion with communications for up to 24 current inputs, (d) 5 I/O configuration with communi-
ware and majority of firmware already work as a transformer or
cations and dual power supply.
small busbar relay in numerous installations.
The new relays have been developed from a simple solu-
munications, and a redundant power supply. tion up towards a sophisticated one. In this way the design was
not biased towards the ability to protect 50+ circuit busbars. As
SAMPLE PROTECTION SYSTEM CONFIGURATIONS a result, the configuration mechanisms, associated software and
Phase-segregated protection schemes, particularly the settings are simple and already known to the user from trans-
ones providing for extra I/O capabilities, equipped with digital former and small busbar applications. The new solutions are
communications means, and supporting multiple zones of pro- easier to engineer as compared with dedicated large-busbar pro-
tection could be configured to protect a variety of busbar con- tection systems.
figurations. Being built on existing hardware and firmware platforms,
Modular hardware platforms such as [7-9] are particular- the new busbar relays are members of existing relay lines. They
ly attractive as they provide for two levels of scalability. First,
each IED could be configured to suit the needs of a given appli-
cation (number of current and voltage inputs, number of digital
inputs, number and type of contact outputs, etc.). Second, the
scheme could be configured from 1, 2, 3, 4, 5 or even more
IEDs depending on complexity of a given application.
Figures 10 through 13 present sample applications.
Figure 10 shows a single-IED protection for a simple
eight-input busbar. The 24 current channels available could be
wired to 8 three-phase inputs; while 3 single-phase zones of
protection could be configured to provide differential protection
for phases A, B and C.
A similar solution for busbars of up to 12 inputs could be
built on 2 IEDs. The first IED uses 2 zones to protect 12-input
busbars in phases A and B. The second IED uses one of its zones
and 12 input signals to protect the remaining phase C.
Using multiple protection systems one could cover large
busbars as long as each section is of a medium size and a check
zone is not required or done externally (Figure 13).
ADVANTAGES AND BENEFITS

The main advantages and benefits of a phase-segregated
approach to digital low-impedance busbars protection schemes
are as follows.
Because the phase-segregated relays are typically devel-
oped based on existing hardware platforms, reusing probably
some 90% of existing firmware, they are much more cost-effi-
cient as compared with dedicated schemes for large busbars.
Being built on existing hardware and firmware, with
thousands of unit-years of field record, the new solutions are
much more mature, and could reach better maturity indices fast,
as compared with dedicated schemes installed in very low vol-
umes. This allows reducing the risk of installing a “new” busbar
relay – the new relays are actually quite mature as their hard- Fig. 11. Sample system configurations: breaker-and-half, two-section busbar with a tie-
breaker, double busbar with a tie-breaker and a transfer bus.
pc_v2 10/19/04 10:24 AM Page 12


devices fails, the system is still operational, providing protec-
tion for all kinds of faults except a SLG fault in the affected
phase. Even more complex systems with communications could
be programmed so that the main functions remain intact when
certain components of the scheme fail.
Oscillography and Sequence of Events recording capa-
bilities are multiplied by using multiple IEDs. Simple software
tools exist to “consolidate” SOE records, and comtrade files into
single files for easier analysis. Programming the IEDs for cross
triggering of various records, and setting up a common clock
reference signal (IRIG-B) allows good synchronization of
records between the IEDs comprising the busbar protection sys-
tem.
User-programmable logic is available in each of the IEDs
allowing easier engineering and more flexible application as
compared with dedicated “hard-coded” schemes that had to be
set up by the vendor, rather than by the user.
APPLICATION CONSIDERATIONS
This section presents application considerations with
respect to protecting complex busbar arrangements where the
objective is to provide for optimum protection by avoiding blind
spots or unnecessary bus outages. As a rule, this task calls for
dynamic adjustments of boundaries of differential zones of pro-
tection, and can be safely accomplished when using numerical
relays.
In the following examples, it is assumed that the dynam-
ic bus replica feature is implemented by configuring a differen-
tial zone using pairs of (current signal, associated connection
status signal). A given current is included in the zone, only if the
Fig. 12. Sample system configurations: single, double and triple busbars. associated status signal is a logic one. In this way, each current
input of the relay could be logically added or removed from the
zone, depending on changing busbar configuration.
share common tools. They could be integrated with other mem-
It is also assumed that the auxiliary logic variables
bers of the relay line. The overall user learning curve could be
required for bus configuration are programmed in user-pro-
significantly reduced.
grammable logic of the relay.
Certain configurations of phase-segregated solutions
exhibit enhanced immunity to relay failures. Consider, for
example, a simple system built on three devices protecting phas-
SWITCHABLE BUS CIRCUITS
es A, B and C without any communications. If one of the Figure 14 presents a sample double busbar arrangement
with bus sections 1 and 2, tie-breaker CB-1, three outgoing cir-
cuits C-1, C-2 and C-3, and a number of CTs and isolator
Fig. 13. Sample system configurations: a large busbar protected by two protection sys- Fig. 14. Sample double-bus arrangement.
tems.
pc_v2 10/19/04 10:24 AM Page 13


switches. It is assumed that the two sections could operate inde- ly connected to the failed breaker. For the busbar in Figure 14
pendently, with the tie-breaker opened or closed. Each section this should be programmed as follows [9]:
could be used as a transfer bus.
Several complex configurations are possible in the bus- FORCE Z1 = BF-1 OR (BF-2 AND S-5) OR (BF-3 AND S-7)
bar of Figure 14. They are addressed in the following subsec- OR (BF-4 AND S-9) (5)
tions. However, if none of the isolators S-5 and S-6, S-7 and S-
8, and S-9 and S-10 could be closed simultaneously, the two FORCE Z2 = BF-1 OR (BF-2 AND S-6) OR (BF-3 AND S-8)
sections could and should be protected independently by two OR (BF-4 AND S-10) (6)
zones of protection. This could be programmed as the following
zone configurations. Assume for example S-5, S-8 and S-9 closed; S-6, S-7
and S-10 opened; and CB-2 failing to trip for a circuit fault. BF-
2 would force Z1 to operate (equation (5)). Z1 would trip break-
TABLE 2. SAMPLE ZONE CONFIGURATION FOR BUSBAR OF
ers CB- 1 (equation (1)) and CB-4 (equation (4)). This would
FIGURE 14. clear the currents towards CB-2, but would leave section 2 and
ZONE 1 ZONE 2 circuit C-2 in service.
Current CT-6 Current CT-5 The above example illustrates a classical situation when
Z1 Input 1 Z2 Input 1
Status On Status On the dynamic bus replica is required. The application may get
Current CT-2 Current CT-2 more complex if certain switching scenarios are allowed in the
Status S-5 Status S-6 substation of Figure 14 (subsection 4.6).
Current CT-3 Current CT-3 Application of the dynamic bus replica is also beneficial
Z1 Input 3 Z2 Input 3 when protecting simple and nonswitchable busbars as explained
Status S-7 Status S-8
Current CT-4 Current CT-4 in subsections 4.2, 4.3 and 4.4.
Status S-9 Status S-10
LINE-SIDE CTS
The settings in Table 2 are interpreted by the relay as fol- A differential zone must be terminated by points where a
lows: the CT-5 current always belongs to Z1; the CT-2 current CT is present in order to locate a fault (selectivity) and a CB is
belongs to Z1 only if switch S-5 is closed; the CT-3 current present in order to isolate the fault (trip action). Although posi-
belongs to Z1 only if switch S-7 is closed, etc. tioned very closely, the two devices have a finite space between
Operation of a given zone should be routed to the break- them. This space is potentially exposed to faults. Depending on
ers using bus configuration at the moment of tripping. In this the mutual position of the CT and CB, the said space may
example, the following trip commands should be created [9]: become either a blind spot or may cause unnecessary trip of the
busbar.
TRIP CB-1 = Z1 OR Z2 (1) Consider an arrangement shown in Figure 15. When the
TRIP CB-2 = (Z1 AND S-5) OR (Z2 AND S-6) (2) CB is opened, the space between the line-side CT and the CB is
TRIP CB-3 = (Z1 AND S-7) OR (Z2 AND S-8) (3) an over-tripping zone: a fault between the CB and CT is not a
TRIP CB-4 = (Z1 AND S-9) OR (Z2 AND S-10) (4) bus fault as the breaker is already opened and no current is sup-
plied from the bus towards the fault. From the metering perspec-
Where Z1 and Z2 stand for operation of differential tive, however, the fault is within the zone. If the current is stat-
zones 1 and 2 respectively. ically assigned to the zone of protection, an unnecessary trip
takes place.
Breaker Failure protection should force an associated dif- Using breaker position as a connection status for the
ferential zone to operate in order to trip all the breakers current- associated current could easily solve the problem. When the
Fig. 15. Line-side CT configuration.

pc_v2 10/19/04 10:24 AM Page 14

Fig. 16. Bus-side CT configuration.
breaker is opened, the current is removed from the zone. As a CB-4 in Figure 14.
result, the zone boundary moves from the CT point to the bus-
side pole of the opened CB. The zone contracts and the unnec- BUS-SIDE CTS
essary trip of the busbar is prevented. As similar situation occurs for bus-side CTs. Consider an
The drop out delay applied to the breaker position signal arrangement shown in Figure 16. A fault between the CB and
is necessary to allow measuring algorithms of the relay to ramp CT is in a blind spot of the bus protection. To clear the fault, the
down after the current is interrupted by the breaker. Otherwise, busbar must be tripped, but the differential zone would not see
a false trip would take place when the breaker opens. this fault.
Refining the example in Figure 1, one should take care of Similarly to the case of the line-side CT, this situation
the spots between CT-5 and CB-1 (for zone 2) and CT-6 and also requires using breaker position as a connection status for
CB-1 (for zone 1). Consequently, the status signals for CT-5 and the associated current. The fault is cleared sequentially. First,
CT-6 would be as follows [9]: protection of the circuit – fed from the CT – responds to the
fault and opens the breaker. When the breaker opens, the CT
Z1 Input 1 Current = CT-6 (7) current is removed from the differential zone. As a result, the
Z1 Input 1 Status = CB-1 + drop out delay (8) zone expands to the bus-side pole of the opened breaker, the
Z2 Input 1 Current = CT-5 (9) fault becomes internal, and the bus protection clears the busbar.
Z2 Input 1 Status = CB-1 + drop out delay (10)
TIE-BREAKER WITH A SINGLE CT
In other words, shortly after CB-1 gets opened, Z1 con- Ideally, two CTs should be used for tie-breakers (see
tracts to the section-1 pole of the opened CB-1, while Z2 con- Figure 14). In some situations, however, a single CT is installed
tracts to the section-2 pole of the CB-1. In this way, Z1 would as shown in Figure 17.
not respond to faults between CT-6 and CB-1, and Z2 would not Two differential zones should be arranged in order to
respond to faults between CT-5 and CB-1. This is desired for provide for selective protection of sections 1 and 2, respective-
optimum selectivity of protection. ly. As a result of having a single measuring point, a fault
The same approach applies to breakers CB-2, CB-3 and
Fig. 17. Tie-breaker with a single CT.

pc_v2 10/19/04 10:24 AM Page 15

COMPLEX SWITCHING STRATEGIES

Complex switching scenarios are often allowed for com-
plex busbar arrangements.
Consider a busbar in Figure 14 and assume a transfer
operation that allows closing the two switches (such as S-5 and
S-6) simultaneously with the tie-breaker closed prior to the
operation. At the moment the two switches are closed, the CT-2
current splits in an unknown proportion between sections 1 and
2. As a result, it becomes impossible to protect sections 1 and 2
separately, and the entire busbar must be protected as one enti-
ty.
First, such a transfer condition must be detected by the
busbar relay. For a busbar of Figure 14, the following auxiliary
flag stands for the condition of the two sections connected via
switches [9]:
A1 = (S-5 AND S-6) OR (S-7 AND S-8) OR (S-9 AND S-10) (11)
Second, the two zones must be re-arranged so that sec-

tions 1 and 2 are protected as one entity. This could be accom-
plished by one of the following:
• Disable Z1 and expand Z2 accordingly.
• Disable Z2 and expand Z1 accordingly.
• Disable both zones and trip from a check zone (if used).
• Expand both zones to enclose the entire busbar.
Fig. 18. End Fault Protection. The last solution is the most elegant because it is sym-
metrical. The following table shows zone configurations that
between the tiebreaker and the CT is internal to Z1, but external account for the transfer scenario.
to Z2. Z1 would trip the tie-breaker, but the fault would not be Assume for example S-5, S-7 and S-10 closed, and S-6,
cleared. S-8 and S-9 opened. Under this condition, the zones are dynam-
Using the position of the tie-breaker for the connection ically adjusted as follows [9]:
status of the CT current for Z2 solves the problem. After TB
opens due to operation of Z1, Z2 expands to the opened break- Z1: (CT-6, CT-2, CT-3) (12)
er, the fault becomes internal to Z2, and Z2 finally clears the Z2: (CT-5, CT-4) (13)
fault by tripping section 2 of the busbar.
When, subsequently, an action is initiated to transfer cir-
END FAULT PROTECTION cuit C-1 from section 1 to section 2, switch S-6 is closed. As a
Consider the arrangement in Figure 18. With the CB-1 result of S-5 and S-6 closed simultaneously, variable A1 is
opened, the bus zone moves from the CT point to the CB-1 asserted (equation (11)). Consequently, the zones are re-adjust-
point, leaving the portion between the CB-1 and CT unprotected as follows (Table 3) [9]:
ed.
A simple overcurrent element responding to the CT cur- Z1: (CT-2, CT-3, CT-4) (14)
rent and activated when the breaker is opened solves the prob- Z2: (CT-2, CT-3, CT-4) (15)
lem. Such End Fault Protection should trip the CB-2 breaker. If
CB-2 is located in the same substation, the application is sim- In other words, both zones protect the entire bus. If a
ple. If CB-2 is located at the remote terminal of a line, Direct fault occurs under this configuration, all the breakers will be
Transfer Tripping capability is required. If DTT is not available, tripped. This may include the tie-breaker (not necessary from
CB-2 would be tripped from the backup protection functions at the fault clearance perspective, but beneficial from the subse-
the remote terminal of the line. quent restoration perspective).
When, subsequently, S-5 gets opened in order to com-
plete the transfer, A1 resets, and the zones re-adjust again as per
TABLE 3. SAMPLE ZONE CONFIGURATION FOR BUSBAR OF
Table 3 [9]:
FIGURE 14 (WITH TRANSFER CAPABILITIES)
ZONE 1 ZONE 2 Z1: (CT-6, CT-3) (16)
Current CT-6 Current CT-5 Z2: (CT-5, CT-2, CT-4) (17)
Status NOT (A1) Status NOT (A1)
Current CT-2 Current CT-2 In this way, CT-2 (circuit C-2) was transferred between
Status S-5 OR A1 Status S-6 OR A1 protection zones 1 and 2.
Current CT-3 Current CT-3 Another operating mode is to transfer all the circuits to
Z1 Input 3 Z2 Input 3 one section, say section 1, keep one circuit on the other section
Status S-7 OR A1 Status S-8 OR A1
Current CT-4 Current CT-4 (say number 2) in order to service the breaker. Once the by-
Z1 Input 4 Z2 Input 4 passing switch is closed, the current of the paralleled CT must
Status S-9 OR A1 Status S-10 OR A1
pc_v2 10/19/04 10:24 AM Page 16


check zone is configured as follows:
Ideally, Z3 should be fed from different cores of the CTs.

There are circumstances when the check zone will not work.
Consider configuration of Figure 19. Because the CT-2 current is
not valid (due to the S-2 switch closed), it is impossible to check
the current balance for the entire busbar. In order to ensure that Z1
would trip for internal faults, the Z3 supervision should be dynam-
ically removed under the circumstances. This could be implement-
ed using the following logic [9]:
Z1 SUPVERVISION = Z3 OR A2 (20)
Z2 SUPVERVISION = Z3 OR A3 (21)
As shown in this section, there are numerous situations

when the boundaries of differential zones of busbar protection
should be dynamically and automatically re-adjusted. Such opera-
tion is natural, easy to program, and safe when using digital low-
Fig. 19. Bus configuration that allows servicing CB-2. impedance relays.
be ignored. The bus section connected to the circuit in question SUMMARY

becomes part of the circuit and is protected by the circuit relay. Several new phase-segregated low-impedance digital pro-
In order to accomplish that protection of the circuit must be tection schemes for medium and large busbars emerged in 2001
transferred to the tie-breaker CT as illustrated in Figure 19. and 2002. Built on existing hardware and firmware platforms these
At the same time, the bus zone that has the non-valid cur- solutions bring in low-cost, initial maturity and simplicity of appli-
rent assigned to it, must be blocked or it may misoperate. For cation.
the busbar in Figure 19 the following auxiliary flags could At the same time, existing busbars are being upgraded to
accomplish this [9]: accommodate new generation and new power equipment. New
substations are built as double-busbars or even more complex
A2 = (S-2 AND S-5) OR (S-3 AND S-7) OR (S-4 AND S-9) (18) arrangements. In new substation designs driven more by cost and
A3 = (S-2 AND S-6) OR (S-3 AND S-8) OR (S-4 AND S-10) (19) less by engineering considerations, motorized switches replace
breakers. Sophisticated switching strategies are allowed to cope
Flag A2 means Z1 is fed with at least one invalid current and with various operational conditions. All this creates major problems
thus it should be blocked. Flag A3 means Z2 is fed with at least one for high-impedance, or nondigital in general, busbar relays.
invalid current and thus it should be blocked. Consider configura- A digital busbar protection scheme is a natural choice for
tion of Figure 19: A2 = 0 (Z1 is not blocked – equation (18), A3 = protecting such complex reconfigurable busbars. Digital relays
1 (Z2 is blocked– equation (19). already reached security of high-impedance schemes [9]. Modern
In this way: relays provide for a sub-cycle tripping time [9]. With the new
• Z1 is reconfigured to protect section 1 by monitoring the fol- phase-segregated approach, the digital schemes become also
lowing currents (CT-6, CT-3, CT-4). affordable and easy to engineer, opening new application opportu-
• Z2 is blocked and would not misoperate. nities.
• Section 2 is protected by protection of C-1 transferred to CT-
5. REFERENCES
[1] Peck D.M., Nygaard B., Wadelius K., “A New
CHECK ZONE Numerical Busbar Protection System with Bay-Oriented
Check zone is a means of achieving extra security. A check Structure”, 5th IEE Developments in Power System Protection
zone is used as a supervising element for regular zones of protec- Conference, 1993, IEE Pub. No.368, pp.228-231.
tion and mitigates problems of wrong status information (incorrect [2] Andow F., Suga N., Murakami Y., Inamura K.,
switch or breaker position recognized by the relay). It will also – at “Microprocessor-Based Busbar Protection Relay”, 5th IEE
least partially – mitigate CT trouble conditions. Developments in Power System Protection Conference, 1993, IEE
The check zone is programmed to monitor the overall cur- Pub. No.368, pp.103-106.
rent balance for the entire busbar. In the example of Figure 14 the [3] Funk H.W., Ziegler G., “Numerical Busbar Protection,
Design and Service Experience”, 5th IEE Developments in Power
TABLE 4. CHECK ZONE CONFIGURATION FOR BUSBAR OF System Protection Conference, 1993, IEE Pub. No.368, pp.131-
FIGURE 14. 134.
ZONE 3 (CHECK ZONE) [4] Evans J.W., Parmella R., Sheahan K.M., Downes J.A.,
Current CT-2 “Conventional and Digital Busbar Protection: A Comparative
Z3 Input 1 Reliability Study”, 5th IEE Developments in Power System
Status On
Current CT-3 Protection Conference, 1993, IEE Pub. No.368, pp.126-130. [5]
Z3 Input 2 Sachdev M.S., Sidhu T.S., Gill H.S., “A Busbar Protection
Status On
Current CT-4 Technique and its Performance During CT Saturation and CT
Z3 Input 3 Ratio-Mismatch”, IEEE Trans. on Power Delivery, Vol.15, No.3,
Status On
pc_v2 10/19/04 10:24 AM Page 17


July 2000, pp.895-901. [8] Kasztenny B., Sevov L., Brunello G., “Digital Low-
[6] Jiali H., Shanshan L., Wang G., Kezunovic M., Impedance Busbar Protection –Review of Principles and
“Implementation of a Distributed Digital Bus Protection System”, Approaches”, Proceedings of the 54th Annual Conference for
IEEE Trans. on Power Delivery, Vol.12, No.4, October 1997, Protective Relay Engineers, College Station, TX, April 3-5, 2001.
pp.1445-1451. Also presented at the 55th Annual Georgia Tech Protective
[7] Pozzuoli M.P., “Meeting the Challenges of the New Relaying, Atlanta, GA, May 2-5, 2001.
Millennium: The Universal Relay”, Texas A&M University [9] B90 Busbar Protection Relay – Instruction Manual. GE
Conference for Protective Relay Engineers, College Station, Texas, publication No.GEK-106241, 2002.
April 5-8, 1999.
APPENDIX
TYPICAL APPLICATIONS OF PHASE-SEGREGATED BUSBAR RELAY [9]
Number of Circuits Number of Breaker Voltage

Checkzone? Comments
Circuits Swichable? Bus Sections Failure? Supervision?
A single IED could be applied. Requires 2 magnetic modules for up to 5 circuits, and
Up to 8 No 1 No No No 3 magnetic modules for up to 8 circuits.
A single IED could be applied. Requires 2 magnetic modules for up to 4 circuits, and
Up to 7 No 1 No Yes No 3 magnetic modules for up to 7 circuits.
A single IED could be applied for basic protection. Requires 2 magnetic modules for
up to 5 circuits, and 3 magnetic modules for up to 8 circuits. Second IED is needed to
Up to 8 Yes 1 Yes Yes No accommodate digital inputs for isolators and breakers, and contact outputs for trip-
ping individual breakers.
Two IED could be applied. First IED protects 2 phases. Second IED protects the third
Up to 12 No 1 No No No phase and allows for extra I/O capabilities.
Up to 24 No 1 No No Yes or No Three IED could be applied. Basic protection is provided for each phase separately.
Up to 24 No Up to 4 No No No Three IED could be applied. Basic protection is provided for each phase separately.
Three IED could be applied for basic protection. Typically 2 or more extra IEDs are
Up to 24 Yes or No Up to 4 Yes or No No No required to accommodate I/Os and BF protection.
Up to 20 Yes or No Up to 4 Yes or No Yes No required to accommodate I/Os and BF protection. Capabilities decrease by 1 circuit
with each bus section for which the voltage must be monitored.
Up to 24 Yes or No Up to 3 Yes or No No Yes required to accommodate I/Os and BF protection. One zone configured as a check
zone. Three zones left for protection of up to three bus sections.
required to accommodate I/Os and BF protection. One zone configured as a check
Up to 20 Yes or No Up to 3 Yes or No Yes Yes zone. Three zones left for protection of up to three bus sections. Capabilities
decrease by 1 circuit with each bus section for which the voltage must be monitored.
pc_v2 10/19/04 10:24 AM Page 18


pc_v2 10/19/04 10:24 AM Page 19

TESTING CONSIDERATIONS FOR MODERN

DISTANCE RELAYS
By Bogdan Kasztenny, GE Multilin
INTRODUCTION self-polarized mho may have directionality problems and need
Protection functions could be designed to reproduce separate directional supervision. Quadrilateral distance func-
widely accepted or even standard operating characteristics, or to tions require an explicit directional line as a part of the polygon.
detect power system faults and other abnormal conditions. Directional supervision may have various forms. The fol-
Time-overcurrent protection (TOC) is a good example of lowing example illustrates a test problem caused by one of
a function that has a widely accepted characteristic. Moreover, them.
its implementation on microprocessor-based relays is standard-
ized so that proper coordination between various TOC devices Example 1
is ensured. Limitations of the TOC function in detecting and dis- A quadrilateral ground distance function has been tested
criminating faults and other conditions are recognized and using single-phase current and three-phase voltage injec-
accepted. Consequently, the function is designed and tested for tion. The angle of the current was controlled to scan
its operating characteristic, not for system faults, unbalance or through 360 degrees of the characteristic. The magnitude of
overload conditions. the current was kept constant, while the voltage was manip-
Distance protection - particularly zone 1 meant for direct ulated in order to find the pickup/dropout points of the
underreaching tripping - is a good example of a protection func- characteristic (“binary search”). Performed on zone 1, the
tion designed to detect and discriminate faults. There is no sin- test returns the expected characteristic. Performed on zone
gle widely accepted operating characteristic for distance protec- 2 it yields an unexpected one (Fig. 1).
tion. Moreover, relay vendors design distance functions for best
possible performance in terms of speed, security and transient
overreach by adding extra comparators, making insides of dis-
tance functions adaptive, etc. The widely accepted designations
of “mho” and “quad” differ between relay models and relay
vendors. Nonetheless, distance functions are tested for their
expected “mho” or “quad” operating characteristics.
Ideally, the two design methodologies converge, making
the task of testing simple. Usually, a conscious effort is made
when designing an advanced protection function to make it
appear straightforward under traditional test conditions, while
providing best possible response to actual power system events.
When the two design perspectives cannot fully converge,
better understanding of the tested relay is required. Adaptive
reactance characteristic is a prime example. Traditional reac-
tance characteristic is a static line perpendicular to the maxi-
mum torque angle line. Such characteristic appears straightfor-
ward when tested using a test set, but would overreach or under-
reach on resistive faults under heavy load conditions. Adaptive
reactance characteristic polarized from either zero-sequence or
negative-sequence current reduces possibility of under/over-
reaching on resistive faults but, when tested, it does not appear
as a static line perpendicular to the line of the maximum torque
angle. The adaptive characteristic does not fit the “expected”
definition of the reactance line resulting in misunderstandings. Fig.1. Tested and expected quad distance characteristic (Example 1).
This paper analyses several practical examples of expect-
ed and actual operating characteristics being at odds. Major pro-
tection functions of a modern distance relay are considered. Figure 2 shows a vector diagram of the voltages and cur-
rents injected during the test. Because the current magnitude
was kept relatively small (it must be kept below the continuous
ADDITIONAL DIRECTIONAL SUPERVISION rating of the relay input), and the reach of the right blinder was
Distance functions use extra directional supervision. The set comparatively high, significant voltage was required to bring
memory-polarized mho comparator provides excellent direc- the apparent impedance to the pickup/dropout point in the area
tional discrimination. After the memory expires, however, the of the right blinder.
pc_v2 10/19/04 10:24 AM Page 20

(1)
Under single-phase current injection:
(2)
The apparent impedance:
(3)
Because the distance element operated when the zero-

Fig.2. Currents and voltages injected during the test (Example 1). sequence directional supervision was satisfied and not when the
right blinder line was crossed, the operating points create an arc
of a circle with the radius of:
(4)
where:
Vnom - is the nominal voltage
IA - is the injected current
Z0 / Z1 - is a zero-sequence compensation setting of the
relay.
All the other points of the characteristic appear correctly
because - given the injected current - the operating voltage was
below the nominal yielding correct zero-sequence direction
(Fig. 4).
Unlike zone 2, zone 1 of the relay from Example 1 does
not use zero-sequence directional supervision - and therefore -
is not affected by this test scenario.
The actual relay design is slightly more complex [1]:
a. The zero-sequence supervision is defaulted to permission
if the zero-sequence voltage is low. Consequently, the
Fig.3. Increased faulty phase voltage causes reversal of the zero-sequence voltage and element would operate even if the zero-sequence voltage
inhibits the tested distance element (Example 1). is reversed, as long as its magnitude is low.
b. The zero-sequence directional supervision is dynamical-
ly removed if a single pole open condition is declared
In particular, for the top right portion of the characteris- during single pole tripping.
tic, the voltage had to be above the nominal in order to reach the c. The zero-sequence supervision circuit has a current-
blinder. During the test, the voltages of the healthy phases were reversal logic built in.
kept at nominal. With the voltage of the faulty phase above The above factors must be considered when testing the
nominal, the zero-sequence voltage got reversed as compared relay. In this particular example, several solutions could be con-
with its natural position (Fig. 3). sidered:
As a result, the built-in zero-sequence directional ele- 1. Apply more current so that the operating voltage is safe-
ment responded to the reverse fault direction and inhibited the ly below the nominal and the zero-sequence voltage is
distance function. In the process of searching for the operating not reversed.
point, the voltage was being reduced. When the voltage dropped 2. Increase the voltages in the healthy phases in order to
below the nominal, the zero-sequence voltage shifted 180 ensure that the zero-sequence voltage is not reversed
degrees to its normal position, releasing the distance function when increasing the faulty phase voltage.
for operation. 3. Remove the zero-sequence directional supervision by
Consequently, the test returned an arc of a circle as cal- forcing open pole conditions.
culated below. One way of increasing the effective current while keep-
Per art of distance protection, the effective input signals ing the individual currents below the continuous rating of the
for the phase-A ground distance unit are: relay is to apply three-phase zero-sequence injection. This must
pc_v2 10/19/04 10:24 AM Page 21


must be above 80% but below 115% of nominal for a certain
period of time in order to establish (and subsequently use) the
memory. In the discussed test, the positive-sequence voltage
was at the level of 33% of nominal. Therefore, the memory was
never validated by the relay nor used. Obviously, being self-
polarized, the tested distance element may pickup during a
close-in reverse fault and/or fail to operate solidly for a close-in
forward fault.
Single-phase injection made the situation worse. By
using the positive-sequence voltage, the relay applies a mixed
cross-phase (positive-sequence) and memory polarization. After
removing the phase-A voltage, all three voltages were zero
yielding no meaningful polarization.
A similar testing mistake is made when injecting sym-
metrical three-phase voltages without considering the nominal
voltage setting (injected voltage too high or too low as com-
pared with the nominal), or injecting the pre-fault voltages for
too short a period of time.
REACTANCE CHARACTERISTIC
Fig.4. Explanation of the incorrect test results (Example 1). The reactance line constitutes the reach-discriminating
boundary of the quadrilateral characteristic. Also, it may be
be approached with caution as the relay may be expecting the used as an extra supervising line for the mho distance function.
negative- and/or positive-sequence currents and may not The latter is beneficial when using an adaptive reactance char-
respond correctly (phase selection [2], overcurrent supervision, acteristic.
etc.). It is a well-known phenomenon that a significant fault
resistance combined with a heavy pre-fault load may appear not
MEMORY AND CROSS-PHASE POLARIZATION as a pure resistance, but may be tilted clockwise making dis-
tance functions overreach, or counterclockwise – making dis-
Distance functions need robust polarization in order to
tance functions underreach. This undesirable effect may be
ensure directional discrimination between close-in forward and
reduced considerably by using the reactance comparator polar-
reverse faults. The actual voltage drops to zero or a very small
ized not from the compensated current (equation 1), but from
value potentially affected by natural errors such as fundamental
either the zero-, or negative-sequence current, or a combination
frequency voltages induced in secondary circuitry, transients or
of the two.
limited relay accuracy. Memory and/or cross-phase polarization
is used to solve the problem.
Example 3
Memory polarization if kept in effect for too long, may
A quadrilateral ground distance function is tested using a
cause a relay to misoperate. A classical example is a power
single-phase injection. The Z0/Z1 ratio is set in the relay to
swing when the signals rotate slowly due to the swing, while the
3.35, 9 degrees. The reactance line seems to be tilted clock-
memorized polarizing quantity remains static. At some point in
wise by approximately 6 degrees from the expected angle
time even under no-fault conditions, the comparators will pick
(Fig. 5).
up, causing a false trip. Therefore, the memory must be in effect
long enough to ride through the breaker fail time for a close-in
reverse fault, and should be dropped after such time in order to
avoid its own problems.
Distance memory logic is not a trivial circuit. It must
address several issues such as how the memory voltage is vali-
dated and invalidated, and when valid, when and for how long
the memory voltage is used. The distance memory circuit must
be understood before testing the relay for memory action.
Example 2
A ground distance function is tested using a single-phase
test set for directional integrity on close-in faults. A pre-
fault voltage is applied in phase A for 20 cycles.
Afterwards, the voltage is changed abruptly to zero, and
simultaneously phase A current is applied in the reverse
direction. The ground distance element in phase A misoper-
ates for such reverse “fault”. When the current is applied in
the forward direction, the element does not operate solidly.
The relay from Example 2 uses positive-sequence mem-

orized voltage for polarization. The positive-sequence voltage Fig.5. Tested and expected quad distance characteristic (Example 3).
pc_v2 10/19/04 10:24 AM Page 22

Fig.7. Illustration of the over/underreach of the static reactance characteristic.
From equation (11) the apparent impedance may be cal-

culated as follows:
Fig.6. Sequence networks for a single-line-to-ground fault fed from one end of the line. (12)
OVERREACH AND UNDERREACH DUE TO FAULT RESISTANCE Equation (12) shows that the apparent impedance is
shifted to the right by the fault resistance. If the two currents,
Consider a single-line-to-ground fault under single-
I0 and IAG, are in phase, the added impedance is a pure resist-
infeed conditions. Fig. 6 presents an equivalent sequence net-
ance and means a horizontal shift to the right. If the compen-
work for this case.
sated relay current, IAG, lags the zero-sequence current, I0, the
In this network:
added value is tilted counterclockwise resulting in possible
underreach. If the compensated relay current leads the zero-
(5)
sequence current, the added value is tilted clockwise, resulting
in possible overreach of the relay (Fig. 7).
and
From equations (12) and (10), the impedance that is
added to the actual fault position defined by Z1 equals:
(6)
(13)
where:
RF - is the actual fault resistance
I0, I1, I2 - are symmetrical currents at the relay location
V0,V1,V2 - are symmetrical voltages at the relay location
The above could be simplified to:
The phase-A voltage and current are in the following
relations to their symmetrical components:
(14)
(7)
(8)
Equation (14) means that even under no-load condi-
tions, a difference between angles of the zero- and positive-
Using equations (5), (7) and (8), equation (6) could be
sequence impedances would tilt the added resistance causing
re-written as follows:
either over- or under-reach of the relay.
(9)
CLASSICAL AND ADAPTIVE REACTANCE CHARACTERISTICS
Classical reactance characteristic is a static line perpendi-
cular to the maximum torque angle line. Using an angle com-
The compensated current is used by any classical parator, the traditional reactance characteristic is implemented
ground distance element: by the following operating equation:
(10) (15)
where ZREACH is the intended reach impedance (reach and

Using equation (10), equation (9) may be now rewritten max torque angle).
as follows: Different version of the reactance characteristic may be
achieved when polarizing from the zero-sequence current:
(11)
(16)
pc_v2 10/19/04 10:24 AM Page 23


establish their pick-up/drop-out state. For example, one particu-
lar ground “mho” function checks the following conditions
before declaring a fault to be within the zone [1, 3].
Dynamic, positive-sequence memory-polarized mho:
IAGZREACH - VA compared with VMEM (18)
Adaptive reactance supervision (to reduce the danger of

overreaching on resistive faults):
IAGZREACH - VA compared with I0ZREACH (19)
Fig.8a. Traditional reactance characteristic. Negative-sequence directional supervision:
I2ZDIR compared with V1MEM (20)
Zero-sequence directional supervision:
I0ZDIR compared with V1MEM (21)
Phase-selection supervision (to inhibit ground distance

functions on double-line-to-ground faults and avoid worsened
accuracy):
I0 compared with I2 (22)
Fig.8b. Zero-sequence polarized reactance characteristic. When memory expires, the V1MEM is substituted by V1.
Attention must be paid to all the comparators when test-
ing the function.
Characteristic (16) provides for an adaptive tilt needed to
overcome the under/overreach effect outlined by equation (14). Example 4
This is illustrated in Fig. 8. A mho ground distance function is tested using single-
The adaptive characteristic (16) preserves constant reach phase injection. The current is kept constant while the volt-
rather than constant reactance. When tested under the expecta- age is being slowly reduced searching for the pick-up/drop-
tion of a constant reactance, the comparator causes some confu- out point. The characteristic seems to be distorted at the top
sion. as shown in Fig. 9.
TESTING THE ADAPTIVE REACTANCE CHARACTERISTIC
The reactance line polarized from the zero-sequence cur-
rent will appear as a skewed line with the amount of tilt depend-
ing on the angle between the zero-sequence current and the
relay compensated current. Even with no pre-fault current, and
single-phase current injection, the line will be titled, depending
on the Z0/Z1 ratio as entered in the relay.
For the data in Example 3, the adaptive reactance charac-
teristic polarized from the zero-sequence current will be titled
by the angle of:
(17)
This is consistent with the results of the test.

Equation (17) holds true for tests with no pre-fault load
and could be applied to calculate the true position of the reac- Fig.9. Ground mho characteristic seems deformed (Example 4).
tance line for traditional tests performed without any pre-fault
load.
In this case, the adaptive reactance tilts clockwise cutting
MULTI-COMPARATOR APPROACHES out a portion of the mho circle as explained in the previous sec-
Distance functions use several conditions in order to tion. Neglecting comparator (19) led to unexpected test results.
pc_v2 10/19/04 10:24 AM Page 24

GROUND DIRECTIONAL OVERCURRENT been tested for directionality. Single-phase injection has
FUNCTIONS been used. The voltage angle has been kept at -180 degrees.
The current angle has been changed in order to scan the 360
Ground directional overcurrent functions are typically degrees of the characteristic. With a 90-degree limit angle
used in conjunction with pilot-aided schemes. The negative- and the maximum torque angle of 86 degrees, the element
sequence or neutral directional functions are fast and sensitive. should operate in the forward direction for current angles
They do not respond to load currents, and enhance resistive cov- between -176 degrees and 4 degrees. The element operates
erage of unit protection [4]. for angles between -185 and 13 degrees when injecting 64V
For better performance ground directional functions are and 5A. If different current is used, the result will change.
often not implemented as straight angle comparators, but use For example, when injecting 10A, the element operates
more sophisticated techniques. Concepts of a positive-sequence between -194 and 22 degrees.
restraint and an offset impedance cause most issues associated
with testing. The tested element uses the concept of an offset imped-
ance [3]. The polarizing voltage is augmented by a small value
POSITIVE-SEQUENCE RESTRAINT proportional to the operating signal:
Example 5
A neutral directional overcurrent function has been tested (26)
for accuracy of pickup. Single-phase injection has been
used. The pick-up value seems to be 6% off compared with
the entered setting. The offset impedance ensures faster and more reliable
operation under low polarizing voltages, i.e. when the local sys-
Ground directional overcurrent functions (neutral and tem is very strong. It also facilitates reliable directional discrim-
negative-sequence) - when used in conjunction with pilot aided ination in series compensated lines. The offset impedance must
schemes - are meant to increase resistive coverage of the protec- not exceed the line impedance and is typically set to a small
tion and, therefore, are typically set very low. If set very sensi- fraction of the latter [1, 3, 4].
tive, these elements may respond to spurious zero- or negative- The offset impedance makes the element more likely to
sequence currents due to natural system unbalances or CT satu- operate, extending the limit angle in both directions. With the
ration. limit angle set to 90 degrees, the extra limit angle added is:
The concept of positive-sequence restraint allows using
sensitive settings while maintaining security [4]. Spurious
ground currents (neutral or negative-sequence) are approxi- (27)
mately proportional to the positive-sequence current. Therefore,
subtracting a small portion of the positive-sequence current
allows fighting the spurious signals. The relay tested in
Example 5 uses the following operating signal for its neutral The higher the current and the offset impedance, the big-
directional overcurrent function [1]: ger the impact on the actual limit angle. The higher the voltage,
the lower the impact.
(23) It was discovered that an offset impedance of 2 ohms was
used in Example 6. Applying the available data to equation (27)
one calculates:
where K is pre-set at 1/16th.
Under single-phase injection:
(24)
The operating signal is precisely 6.25% lower than

expected explaining the test results in Example 5. Under pure
zero-sequence injection (all three currents in phase), the follow-
ing holds true:
(25)
yielding expected results. In this case an intentional rela-

tion between sequence currents has been introduced in the
design of the element. The positive-sequence current lowers the
effective operating signal. Neglecting this fact affects testing.
OFFSET IMPEDANCE
Example 6
A negative-sequence directional overcurrent function has
Fig.10. Impact of the offset impedance on the actual limit angle (Example 6).
pc_v2 10/19/04 10:24 AM Page 25


Under 5A injection, the actual limit of operation are valid. Actually, due to inappropriate signal injection, the relay
–176-9 =-185 deg, and 4+9=13 deg. was tested with frequency tracking effectively disabled.
Given equation (28), a valid test requires feeding the
relay with appropriate amount of positive-sequence voltage.
8. TRANSIENT TESTING
Modern test equipment is capable of waveform playback.
Under 10A injection, the actual limit of operation are Microprocessor-based relays are capable of recording faults and
–176-18 = -194 degrees, and 4+18=22 degrees. other disturbances. More and more often actual fault records are
This explains test results of Example 6. played back in order to verify relay design and/or settings.
The available fault records, the playback equipment, and
OFF-NOMINAL FREQUENCIES the relay under test must be used appropriately in order to yield
Example 7 meaningful test results.
Distance, voltage and current protection elements have For example, typical playback equipment accepts wave-
been tested for accuracy under off-nominal frequencies. forms that are evenly spaced in time (constant sampling fre-
The relay seems to be affected by off-nominal frequencies. quency). A typical microprocessor-based relay uses frequency
Figure 11 presents test results. An error of 1-2% is meas- tracking to compensate for off-nominal frequencies. As such,
ured per each Hertz of difference between the actual and the relay would produce a record of variable sampling frequen-
nominal frequencies. cy. When played back, such a record will not represent the event
correctly. Prior to the playback, the record should be re-sampled
to a constant sampling rate to suit the playback equipment.
Generally, care must be taken to make sure the playback
equipment does not alter the record in a way that invalidates the
test. One example of such alternation follows.
Example 8
A fault record has been played back to the relay (Figures 12
and 13). It is an internal C-to-ground fault, followed by an
external A-to-ground fault. A single-pole-tripping relay that
produced the record correctly tripped pole C and later it
overreached for the external fault and tripped poles A and B
of the 500kV, 250-mile line. The test is to verify both the
reach setting and accuracy of a different relay model. The
relay under test was consistently overreaching for the sec-
ond external fault.
Fig.11. Accuracy of pickup and reach under off-nominal frequencies (Example 7). It turns out that the playback equipment in Example 8
accepts a limited number of samples. For large records, the soft-
ware cuts the tail part of the record (the original record is shown
A frequency tracking algorithm (or alternatively frequen- in Figures 14 and 15). In addition, the software analyzes each
cy compensation) adjusts the sampling frequency of a micro- waveform scanning it from the end of the truncated record, and
processor-based relay to the actual power system frequency in artificially zeros the waveform from the end of the record to the
order to keep the number of samples per power cycle constant, nearest zero crossing. This is done in order to avoid overvolt-
and by doing so, ensure accurate digital estimation of currents ages due to currents being chopped. The operation is performed
and voltages. for currents and voltages independently in each channel.
Distance relays may use sophisticated frequency tracking As a result, in the case of Example 8, the voltage is
algorithms. This is particularly true for single-pole tripping removed well before the current. For approximately half a
relays where no single voltage is a good choice of a frequency- cycle, the voltage is zero while the current remains high (Fig.
tracking signal. During single-pole tripping particular phases 16). This is causing an operation of the distance function regard-
may get de-energized and their voltages may get severely dis- less of the set reach. Proper test should include actual switch-off
torted by transients related to shunt reactors. transient as shown in Fig. 14 and 15, or disregard the switch-off
The relay tested in Example 7 uses the Clarke transform transient. An artificial switch-off created by the test equipment
to derive an effective input signal for frequency tracking [1]: is a root cause of the problem.
(28) SUMMARY
Modern distance relays are designed to ensure best pos-
The frequency signal will zero-out under pure zero- sible performance under actual power system conditions and not
sequence injection preventing the relay from measuring and necessarily to follow any standard operating characteristics.
tracking the frequency. It was discovered that the relay in New concepts are being introduced in order to improve
Example 7 was fed with three voltages being of equal magni- response to system faults and other abnormal conditions. Even
tudes and in phase. As a result, the effective signal zeroed-out simple and well-known protection functions may follow more
and the relay did not track frequency. Test of Example 7 was not sophisticated design philosophies than indicated by their verbal
pc_v2 10/19/04 10:24 AM Page 26

Fig.12. Voltages used in the playback (Example 8). Fig.15. Original currents recorded in the system (Example 8).
Fig.13. Currents used in the playback (Example 8). Fig.16. Voltages used in the playback (Example 8).
test.
Ultimately, advanced relays will have to be tested either
using detailed power system models (steady-state or transient)
or for their own design equations.
REFERENCES
[1] D60 Line Distance Relay – Instruction Manual. GE
Publication No. GEK-106341, 2003.
[2] Kasztenny B., Campbell B., Mazereeuw J., “Phase
Selection For Single-Pole Tripping – Weak Infeed Conditions
And Cross Country Faults”, ”, Proceedings of the 27th Annual
Western Protective Relay Conference, Spokane, WA, October
24-26, 2000.
[3] Kasztenny B., “Distance Protection Of Series
Compensated Lines – Problems And Solutions”, Proceedings of
Fig.14. Original voltages recorded in the system (Example 8). the 27th Annual Western Protective Relay Conference,
Spokane, WA, October 22-25, 2001. Also Presented At VI
Simposio "Iberoamericano Sobre Proteccion De Sistemas
designations. Electricos De Potencia", Monterey, Mexico, November 17-20,
Too often functions of advanced relays are tested for 2002.
operating characteristics extrapolated from their names or ANSI [4] Kasztenny B., Sharples D., Campbell B., Pozzuoli
numbers disregarding their actual design equations. M., “Fast Ground Directional Overcurrent Protection–
Advances in protective relaying techniques and relays Limitations And Solutions”, Proceedings of the 27th Annual
call for testing procedures that are either closer to actual power Western Protective Relay Conference, Spokane, WA, October
system conditions, or follow design philosophies of relays under 24-26, 2000.
pc_v2 10/19/04 10:24 AM Page 27

VECTOR JUMP RELAYING:

An Economical Solution for Distributed
Generation Islanding Protection
By Arvind Chaudhary, Cooper Power Systems
INTRODUCTION One concern with this technique is that the communica-
The use of Distributed Generation (DG) in distribution tion channel and equipment to implement the transfer trip can
power systems has increased over the last decade, spurred by be expensive. Another concern is the dependency on this com-
the ability to defer the building of new substations and by new munication channel to implement the trip. Another technique
cost-effective technologies (fuel cells, micro-turbines, etc.). that can be used for the detection of islanding of the local power
Interconnecting dispersed generation to the utility system system is the use of a vector jump relay. This type of relay does
requires robust relaying to detect the loss of the utility system not require communications equipment and is comparatively
and to take preventive measures. One critical measure is to dis- economical.
connect the local generation from the utility system (intentional A vector jump is the step change in apparent frequency
islanding) before the utility tie recloses. This islanding avoids that occurs when a distributed generator and its local load are
damage to generator shafts. The islanding process can also suddenly disconnected or unintentionally islanded from the util-
include selective disconnection of less important local loads, to ity. The vector jump relay detects a sudden change in the direc-
ensure that any important process loads can continue to be sup- tion of power flow by examining the system voltage signal. The
plied by the local generation. relay’s operation is made more secure and dependable by
Figure 1 shows a typical industrial DG configuration. adding voltage and frequency conditions to the vector jump.
Breaker 1 is owned and operated by the utility. Breaker 2 is
owned and operated by the DG owner. One of the key relaying ISLANDING DETECTION USING VECTOR JUMP
requirements is to prohibit the reclosing actions of Breaker 1 for RELAY
a fault on the line between Breaker 1 and the point of common In the case of a generator operating in parallel with a util-
coupling. The relay must stop acting as a source, and thereby ity distribution network, when the utility network is suddenly
energizing the Area Electric Power System (EPS) to which it is disconnected (Breaker 1 of Fig. 1 opens), the generator becomes
connected, prior to reclosure by the Area EPS. Traditionally, a islanded and is left to supply all of the remaining connected
leased communication line is used to provide a transfer trip sig- load, including the share of load formerly supplied by the net-
nal to Breaker 2 the moment Breaker 1 opens in response to a work. The vector jump relay includes an element designed to
fault. Since Breaker 2 is now open (the local DG unit is island- detect the disturbance produced by the sudden variation of the
ed), the reclosure sequence of Breaker 1 can now take place load at the terminals of a generator. This disturbance results in a
without the possibility of any damage to the distributed genera- proportional variation of the angular displacement between the
tor. generator’s electromotive force E and the voltage at its termi-
nals or point of relaying, V (Fig. 2 and 3).
The resulting ∆a of the angle a is referred to as a vector
jump or vector surge. If the circuit breaker connecting the
islanded generator to the rest of the network is automatically
reclosed, it is possible that the voltage displacement between the
generator’s bus and the network can be too large for safe paral-
leling of the generators. Rapid detection of this condition pro-
vides the opportunity to:
• Trip the utility tie breaker (Breaker 2) or the generator’s cir-
cuit breaker (Breaker 3….in Figure 1????) before the
reclosing of the tie utility’s breaker occurs, thus avoiding
possible serious damage to the generator itself.
• After tripping Breaker 2, it is possible to trip any non-criti-
cal local load in order to match the generator’s output to
that of any critical processes in an industrial environment.
Such a relay can detect a vector jump in a 2° to 30° range,
resulting in a trip signal being issued in less than three cycles
Local Electric Power System
(Fig. 4). One of Cooper Power Systems’ relays uses a moving
average of the last five cycles’ periods as a comparison value in
Fig.1. Interconnection diagram with definition of terms order to reduce spurious operations. At each cycle, a compari-
pc_v2 10/19/04 10:24 AM Page 28

~U = E – V = I1 • JZd ~U' = E – V' = IL • JZd = ( I1 + I2 ) JZd
Fig. 2. Normal a between generator and measured voltages Fig. 3.Resulting ∆a between generator and measured voltages
son is done between the average of the last five cycles’ periods SETTING OF THE VECTOR JUMP FUNCTION
to the average of the last five cycles’ periods measured at the The definitions for relay dependability and security are:
end of the previous cycle. The detection of the disturbance can • Dependability is defined as the measure of certainty that the
be made in either single-phase or three-phase modes. relays will operate correctly for all faults for which they are
• In the single-phase mode, tripping takes place as soon as a designed to operate.
measured ∆a exceeding the set level is detected on one of • Security is defined as the measure of the certainty that the
the three-phase voltages. relays will not operate incorrectly for any fault (e.g., a fault
• In the three-phase mode, tripping takes place only if the outside the intended zone of protection).
value of ∆a above the set level is detected on all three phas- The vector jump relay must be set to pickup when the
es at the same time. utility main feed is disconnected. It is essential that there is an
Single-phase mode is more sensitive than three-phase interchange of power occurring at this time. If there is no inter-
mode in detecting vector jump conditions; however, it is also change of power, it is impossible to detect the loss of the utility
more sensitive to spurious disturbances. An undervoltage ele- main feed. This is dependability. In order to be certain that the
ment is used to block the vector jump function if the voltage relay will detect the system’s vector jump, the setting for the
drops below a user adjustable level threshold. A contact input vector jump should be set lower than the expected vector jump
operated by a normally open auxiliary contact 52a of a circuit that occurs during the minimum power interchange.
breaker blocks the vector jump functions when the circuit The vector jump relay has to be set such that it does not
breaker is open, as well as for a period of time after closing the operate for an external fault on the system and also does not
breaker. operate when the largest load is being picked up by the local
The value of ∆a as a function of the power variation of generator. Ideally, the vector jump relay should not pickup for a
the generator (∆P) in percentage units of its rating as it passes fault on the internal system. This is security. These settings for
from the normal situation to an islanded situation may be the vector jump should be higher than the vector jumps these
approximated as: conditions create to avoid unwanted operation with certainty.
Security of the relay can be enhanced by using the three-phase
∆a= (0.3 – 0.4) ∆P mode in preference to the single-phase mode. The vector jump
function can be ANDed with underfrequency AND undervolt-
Where the 0.3 multiple is generally more applicable to age AND (.not. negative-sequence voltage) AND (.not. zero-
large generators and the 0.4 multiple to small generators. The sequence voltage) elements for even greater security.
use of a computer-based transients analysis program can be used It is recommended that the vector jump relay voltage
to more accurately determine the ∆a setting in order to avoid input be fed from voltage transformers located at the local gen-
spurious trips. erator terminals. This location will provide larger vector jump
magnitudes and the ability to trip the local tie breaker (Breaker
2 in Fig. 1) when the main utility feed is lost. The 52a contact
of local tie breaker is required as an input to the relay. This 52a
contact input is used to block the vector jump functions when
the circuit breaker is open as well as for a period of time after
closing the tie breaker with an appropriate synch-check func-
tion.
Information needed to provide the settings:
1. The rating of the local generator.
2. The minimum interchange power condition, either to the
utility or from the utility.
3. The change in the generator’s terminal voltage angle (∆a)
when the utility main is lost at minimum interchange
power. Also, the change in frequency and in voltage under
Fig. 4. Effect on system voltage as seen by relay due to vector surge this condition.
pc_v2 10/19/04 10:24 AM Page 29


4. The change in the generator’s terminal voltage angle (∆a) local area power system for the detection of the loss of the util-
when an external three-phase fault occurs on the main utility system. If both the local area and utility are not interchang-
ity and is cleared by the main utility relaying system. Also, ing any power and are in synchronism (floating) and either
the change in frequency and in voltage under this condition. Breaker 1 or Breaker 2 opens, then detection by this method is
Note that three-phase fault condition is specified, implicit- not possible.
ly assuming use of the three-phase mode of the relay.
5. The change in the generator’s terminal voltage angle (∆a) REFERENCES
when an internal three-phase fault occurs on the local sys- 1. IEEE Standard 1547-2003™, “IEEE Standard for
tem and is cleared by the local system relaying. Also, the Interconnecting Distributed Resources with Electric Power
change in frequency and in voltage under this condition. Systems”.
Note that three-phase fault condition is specified, implicit- 2. “Protection Relays for Mains Failure with respect to
ly assuming use of the three-phase mode of the relay. G59”, J. C. Russell, ERA Conference on Protecting Electrical
Networks and Quality of Supply in a Deregulated Industry,
Calculations needed to provide the settings: London, UK, pp. 7.3.1-7.3.10, 1995.
Note that it is not possible to detect the loss of the utility 3. “A New Microprocessor-Based Islanding Protection
main when the net interchange of power is zero or extremely Algorithm for Dispersed Storage and Generation Units”, M. A.
low. The minimum available setting for ∆a is 2°, hence the min- Redfern, J. I. Barrett, and O. Usta, IEEE Transactions on Power
imum amount of interchange power should be four percent of Delivery, Vol. 10, No 3, July 1995, pp. 1249-1254.
the local generator’s rating by employing other characteristics, 4. Vector Jump relay manual, S150-23-1, January 2000,
such as under- or overvoltage, and under- or overfrequency, Cooper Power Systems.
with suitable time delays and then logically ANDing these char- 5. Setting the Vector Jump Relay, R150-23-2, March
acteristics with the vector jump function. 2000, Cooper Power Systems.
6. Distribution Feeder Protection and Monitoring Relay
ADVANTAGES, DISADVANTAGES, AND FUTURE iDP-210, 165-210, March 2002, Cooper Power Systems.
IMPROVEMENTS
The main advantage of this method of detection of Arvind Chaudhary (S’85–M’85–SM’94) is a Staff Engineer with
islanding operation is economy over the traditional communica- the Protective Relays and Integrated Systems Group, Cooper
tion-based methods. Power Systems, South Milwaukee, WI. He is responsible for
The main disadvantage of this method is that there must relay applications for the Cooper line of relays and relay set-
be a sufficient power flow between the utility system and the tings for power system equipment.
pc_v2 10/19/04 10:24 AM Page 30


pc_v2 10/19/04 10:24 AM Page 31

ASSET MANAGEMENT:
Purchasing Data Integration for Relay
Protection
By Joseph Stevenson, ENOSERV LLC
We've moved skillfully into the 21st century, when at the tainly been beneficial. Gone are many of the manual processes
close of the 20th we thought we could see the end-all of catas- of the past; in the wake however, is a whole new set of redun-
trophes because of Y2K. Moving ahead in this post-9/11 world, dant, time-consuming steps necessary to maintain the data
many of us were preoccupied until an incredible coincidence exchange between automated systems, particularly so, wherev-
last August reminded us once again how tenuous our relation- er software meets hardware in terms of computerized equipment
ship with technology can be. Going forward, realistic circum- and the data exchange between assets. Typically, there is no
stances must include examination of even the fantastical if the communication between products from different manufacturers
stewards of the nation's electric grid are to ensure its reliability, and so O&M personnel are shackled to projects of creating use-
especially now with control spread widely among so many dif- ful information exchanges between assets which, in turn, takes
ferent stakeholders. Gone are the "good 'ol days" in many ways, their attention and time away from their focus of maintaining
indeed. system reliability.
None of this is news; but the story here is how the power Take relay testing as an example. Relay technicians and
industry - affecting virtually all aspects of our lives today, with engineers represent a core human asset for ensuring system reli-
an infrastructure on life support and a narrowing, downsized ability. At their disposal are the relays themselves which, today,
and aging workforce - can adapt. Automation technology might are still by proportion largely electromechanical but are increas-
help, but companies now are somewhat wary; they've bought ingly being replaced with microprocessor-based relays.
software and computerized equipment in the past and upgrading Microprocessor relays perform several protection functions pre-
from what they've already implemented and established as sys- viously impossible from just one device, bringing with them a
tem-wide practice creates political and logistical problems. whole new level of functionality and complexity. Testing to
Besides, new technology costs money
and the payoff takes time - neither of "Linking the overall computerized asset management system down to the
which is in great supply these days. relay in the field…is the major breakthrough in system reliability since the intro-
So what is the solution? For many duction of the microprocessor relay and automated relay testing."
companies, it's integration.
Since the dawn of the first
chipped wheel, manufacturers have
done things differently. Products com-
peting in the market ever since give
the consumer perceived benefits in
terms of design, function, price, etc.,
of one make over another. By now in
our ever-widening economy, particu-
larly as applied in the electric power
industry, choice is present but con-
spicuously absent in the view of many
power companies is a universal solu-
tion when it comes to automation. For
those companies the belief is that,
given the complexity of the processes
involved in delivering the product,
there simply isn't one "off-the-shelf"
software solution. The power compa-
nies then take on the task themselves
of trying to integrate their disparate
systems into what they want with,
again, fewer human and financial
resources.
In this new era of microproces-
sor technology, automation has cer-
pc_v2 10/19/04 10:24 AM Page 32


deciphering paperwork as a neces-
sary step in the whole testing process.
Today, the process is different.
With automated testing, the laptop
computer and the various automated
relay testing software systems [like
AVTS (Megger), RTS (ENOSERV) ,
ProTest (Doble), MasterTest
(Manta), Omicron, et al] have
replaced the labor-intensive process-
es of the paper system. Now, histori-
cal records and settings can be
backed up to a database affording
both engineers and technicians more
time to get on with their core func-
tions involved in maintaining system
reliability. The advent of the micro-
processor relay made automated
relay testing compulsory though, and
so both engineers and technicians
alike had to learn the new technology
to keep up while the industry entered
the digital age.
Automated testing and the
microprocessor-based relay have
been around for over a decade now
ensure they're programmed properly with the correct settings, and while advances surrounding the
and that they're used in the proper applications, raises the impor- whole process of relay testing have emerged and have been
tance of the scope of work that protection engineers and techni- widely incorporated into reliability maintenance throughout the
cians do. industrialized world, there still remains several critical prob-
Testing relays requires the use of a secondary injection lems. With more demand than ever on technicians and engineers
unit, or test set, that applies currents and voltages into the relay to be knowledgeable in computer science, and with much (if not
from which the relay technician can make a determination on its all) of that training only available on-the-job, the need to main-
condition. The results of the test get recorded and if the relay tain a capable workforce comes at a price… in real dollars.
requires adjustment or settings changes, then the technician cal- Asset managers must leverage human, financial, and physical
ibrates or programs the relay accordingly. The technician must assets all the time. The problem is finding the resources to keep
be knowledgeable on power, relays, the test set, and the settings personnel technologically knowledgeable in a fast-paced cli-
as applied to the relay and what they mean in order to perform mate; the challenge is cultivating knowledge and knowing
tests and maintenance confidently. Likewise, the protection where it counts most.
engineer who initiated the settings must also use his or her back- Going further, there's a data gap between disparate soft-
ground in power and apply it to the scheme within which the ware systems. Even computer-savvy personnel hit dead ends
protective relay functions in order to create settings that make trying to get different systems to talk. In the case of relay test-
sense. ing, each test set manufacturer has its own software and with a
Straightforward as the process of relay testing sounds, crew of technicians using test sets from different manufacturers,
the critical component to it is the results; or rather, how the test the test settings, test results, etc., must be backed up to different
data gets turned into useful information about the condition of databases in different ways. The protection engineer is left to
the system's reliability. Historically, when testing electro- sort through data from different locations to assess the condition
mechanical relays manually for example, the technician pow- of the electric system under his watch.
ered up the test set and hand-calculated values expecting the Companies have spent millions working on in-house sys-
relay to operate within that pre-determined range. If it did not, tems to integrate their software resources, and largely those
it got adjusted or pulled out of service. What guided the whole applications provide only a rudimentary way of exchanging
testing process were the paper setting sheets from the engineer data. The whole point of automation is to eliminate manual
and the paper test forms and reports the technician completed steps and create useful information from data. The big problem
during and after testing. Following the testing process, the of software in this industry shows itself at those times when the
reports got filed back at the office, probably with the setting different systems must be "tweaked" in order to make them use-
sheets. The test forms stayed out in the field at the substation. ful to personnel who have come to rely on them in order to get
With the paper system, engineers reviewed reports and their work done.
setting sheets line-by-line, page-by-page. Hours could turn into In this time of constrained budgets and fewer human
days spent on reviewing the records, particularly whenever an resources, outsourcing data integration to third-party software
error (like a transposition) got detected and correcting it meant companies pays off. Companies taking advantage of "off-the-
revisiting each record affected. Technicians also labored shelf" applications like Cascade (Digital Inspections), Maximo
through the manual system in the field spending valuable time (MRO), and PowerBase (ENOSERV) avoid the financial pit-
pc_v2 10/19/04 10:24 AM Page 33


falls widely associated with homegrown systems. In terms of PowerBase, even communicates with older test equipment,
relaying, solutions currently exist in the marketplace to maintain lengthening existing asset lifespans, keeping tools relevant and
and track all relay data from the device in the field to the over- useful. Saving/making time and money is what automation is all
all, enterprise-wide asset management system. Relay techni- about; the loaded benefit is any system that is easy to use and
cians and engineers have historically been isolated in the sense can bridge any data gaps.
that the records they must maintain wind up in their own loca- Leveraging asset maintenance with existing resources is
tion on the server that only a handful of people even know the ongoing challenge facing the electric power industry. At the
about. Linking the overall computerized asset management sys- heart of the issue is reliability, but today's industry is wrought
tem down to the relay in the field, and automating all processes with organizational and political characteristics not seen before
associated with testing and maintenance in between, is the in history. What effect will new policy have on the grid? All that
major breakthrough in system reliability since the introduction is known is that there is a certain threshold the industry is
of the microprocessor relay and automated relay testing. approaching and beyond that, it's essentially anyone's guess as
Companies taking advantage of commercially available to how the industry will respond to ever-increasing demand and
integration solutions remedy problems of time constraints and costs with a labor pool that is ever-shrinking. Today, at least in
benefit by seeing return on investment contribute to long-term relay protection, the solutions are available for purchase in the
profitability. Available integration software, like RTS and marketplace.
pc_v2 10/19/04 10:24 AM Page 34


pc_v2 10/19/04 10:24 AM Page 35

PROTECTION AND CONTROL UPGRADE -

USER ADDED VALUE
By Schneider Electric
INTRODUCTION USER BENEFITS
Providing power to customers is no longer sufficient. To achieve the benefits of protection and control
Industry and utilities must deliver power safely, reliably and upgrade, we need to:
cost effectively. The devices that control and protect electrical • Understand which electrical assets are operating most effi-
assets must ensure the safety of people and property, improve ciently,
continuity of service and contribute to the operation and main- • Track performance and update the energy usage model,
tenance of the plant. Industry and utilities recognize that the • Know if all facilities are operating using best practices and
power system components, such as cables, transformers, break- new technology, and,
ers and control devices require more visibility. These assets • Identify opportunities for process improvement.
need to be protected, monitored and maintained. The driving The user benefits are threefold: operational, economic
questions are: and maintenance-related. Several examples show the benefits
• How do you improve equipment reliability to minimize users can derive from a protection and control upgrade.
“down time”? Operational benefits include Analysis and Control.
• What actions can you take in the near future and long term?
• How do you increase operating equipment efficiency? ANALYSIS
The protection and control engineer is concerned with By tracking critical variables over extended periods of
limiting the effects of disturbances in a system network, which, time and performing quick graphical analysis of relevant data,
if they persist, may damage the plant and interrupt the supply of plant engineers can identify abnormal operating conditions.
energy. These disturbances, described as faults (short circuits This includes:
and open circuits) or power swings, result from natural hazards, • Fault recording/failure analysis: (using diagnostics to deter-
plant failure or human error. False relay trips are a contributing mine phase unbalance, tripping currents, and alarm history
factor in several of the system disturbances. As we identify crit- occurring on the electrical network),
ical circuits or equipment, we must also assume that false trips • Creating custom quantities: by incorporating breaker con-
of these critical circuits can have as large or a larger impact trol and indication,
upon the system as does a failure to trip. The implementation of • Programming the front panel to display detailed alarms.
digital protection will increase the functionality and flexibility
of medium- and low-voltage switchgear operations. In return, CONTROL
improved asset operation will result from using system monitor- Correcting abnormal operating conditions will improve
ing and control to plan electrical asset enhancements and the production efficiency and reduce operational inefficiencies.
changes. The relevant functions are:
Electromechanical and first generation solid state relays • Breaker opening by protection and by remote control,
do not allow auto testing of internal circuits or provide an alarm • Breaker closing by remote control,
in the event that a failure is detected. Electromechanical relays • Trip circuit supervision.
have many mechanical parts, which may become clogged with
dirt or corroded due to environmental conditions affecting the ECONOMIC BENEFITS
operation, calibration and movement of disks. Most of the first
generation solid state relays employ series, shunt and switched Predictive maintenance is becoming extremely important
mode power supply designs. For a variety of reasons, if these in the efforts of industry and utilities to deal with reduced per-
power supplies fail, the measuring circuits are inoperative and sonnel while meeting customer demands. Information generat-
no protection is available to the network. Most of the solid-state ed from a protective unit can help avoid unnecessary capital
relays in use do not have the means to detect the failure and ini- investment. Modern protective devices measure and calculate
tiate an alarm. Electromechanical relays can be in an unrecog- numerous monitoring functions and allow event driven mainte-
nized defective state for a long time (until a scheduled test), but nance, such as:
a digital relay self-test will inform you of a failure immediately. • Trip circuit supervision to detect fault wiring between the
The different levels of reliability between the metering devices, protection and switchgear trip units,
the communication system and the single-phase relays can also • Operating time to measure the breaking device’s opening
be a problem, because the reliability of the protective system and closing times in milliseconds,
depends on its weakest part of many devices in the electrical • Charging time to measure the breaking device control
network. charging time, where control charging is done by an elec-
tric motor as soon as the circuit breaker is closed,
• Cumulative breaking current to count according to 5 levels
pc_v2 10/19/04 10:24 AM Page 36


of fault current,
• SF6 pressure switch: to provide an alarm,
• Self-test to assess relay health,
• Logical discrimination to achieve better electrical network
coordination.
This information will facilitate switchgear maintenance
and maximize the visibility of electrical assets thereby increas-
ing the equipment’s operating effectiveness.
MAINTENANCE BENEFITS
Maintenance benefits include integration and informa-
tion
INTEGRATION
By integrating protection, metering, communication,
control and monitoring capabilities, digital relays significantly
reduce wiring and thereby increase reliability. Spare parts
inventory is also minimized.
INFORMATION
Information allows multiple users to access critical data
when and where they need it. This helps improve productivity
and reduce administrative costs. Plant personnel can easily gen-
erate custom data reports.
With a history of supplying electric protection and con-
trol products, Schneider Canada Services has the expertise to
upgrade existing systems with the most modern components to Fig.1. MV Protection Upgrade with SEPAM
optimize the performance of medium-voltage and low-voltage
protection relays. A medium voltage protection upgrade is The intelligence of the switchgear is in its protection and
shown in Fig. 1. control equipment. Upgrading both medium- and low-voltage
From the system upgrade, industrials and utilities gain: protection units with new protection and control devices can
• no investment in infrastructure, significantly extend the life of the switchgear at a significantly
• short implementation times, lower cost than purchasing and installing new switchgear. The
• minimal downtimes to operation, and, user also gets increased electrical system performance data and
• execution time by cost-effective authorization proce- reliability with minimal impact on the physical installation.
dures.
Recent studies show relay data is used in more opera- For more information, contact Pratap Revuru, P. Eng., Manager,
tional and engineering applications today than ever before. Market Research and Innovation, Schneider Canada Services,
Remote engineering workstations now use oscillographic event at pratap.revuru@ca.schneider-electric.com or (905) 678-7000.
reports to assist in locating faults and understanding the type of
faults.
pc_v2 10/19/04 10:24 AM Page 37

ANALYZE RELAY FAULT DATA

TO IMPROVE SERVICE RELIABILITY
By Roy Moxley, Schweitzer Engineering Laboratories, Inc.
ABSTRACT protection, its speed of operation, and its impact on the overall
Protective relays are at the core of maintaining electric power system control system. In examining the success or fail-
service to as many customers as possible during a system distur- ure of relays to perform to these measurements, choices can be
bance. For this reason, an improvement in protection quality is made that improve, rather than degrade, the overall relay system
directly reflected as an improvement in customer service. To performance.
improve protection quality it is useful to identify and record per- Analysis of false operations, failure to operates, speed of
formance measures that can be evaluated for system impact. operation, and general system operation make evident statistical
Using 18 months of data (January 1996 – August 1997), and anecdotal information. This information supports possible
detailing every relay operation on an anonymous utility system changes in protective practices that can reasonably improve sys-
(1400 operations), this paper analyzes the faults and protective tem protection and operation.
system operation to determine relay operation and performance
of electromechanical, solid-state, and digital relays. Specific STATISTICAL MEASURES
system quality measurements include relay misoperations, relay Making a first pass of data from the selected utilities’
failures to operate, relay delayed operations, and accessory relay operations, it can be grouped as follows:
component failure, such as fault recorder, trip circuit, communi- • 1425 total events
cations system, or targeting system. Results are compared to the - 1346 correct operations (94.5%)
values given by Working Group I17, Transmission Relay - 66 incorrect operations (4.6%)
System Performance Comparison for 2000 and 2001 [1]. - 13 failures to operate (0.91%)
For each protection quality measurement, examples of The IEEE Power System Relay Committee Working
negative responses and their power system impact are present- Group I17 Report, Transmission Relay System Performance
ed, as well as statistical values detailing the probability of a neg- Comparison [1], further breaks down statistical performance by
ative response. In conclusion, a discussion of reasonable mitiga- voltage class and type of relay operation failure. The way the
tion techniques and their cost and benefits, based on the data, is subject utility would report their actual faults to the working
presented. group is shown in Table 1.
While these ratios are significant in making a relative
INTRODUCTION measure of overall system performance, the individual events
Looking at actual system events provides fascinating that make up each category better show how improvements in
details in the life of a power system. A massive explosion at an protective system design can be made.
industrial facility breaks windows a half-mile away and causes
electromechanical contacts to bounce closed. Pelicans fly into a SPECIFIC MISOPERATIONS (SEE APPENDIX 1)
sub-transmission line and relays clear the fault after a time Even examining the reasonable percentage of misopera-
delay. Individual events are significant in that they form part of tions (less than 5%) is difficult due to the large number of
a pattern; the analysis of which can be used to improve relay events. It can be made more manageable if the misoperations
operation. are divided into groups, based on cause. A quick read of the fail-
The power system data used in this paper includes ures makes a few broad categories quickly apparent.
electromechanical relays, solid-state
relays and microprocessor relays. The TABLE 1 IEEE WORKING GROUP I17 INCORRECT OPERATION REPORTING
utility involved has an excellent serv- % Incorrect Operations (due to relays), Years 1996-1997
ice record for its customers and is Voltage (Kv) Failure Failure to Slow Trip Unnecessary Unnecessary Failure to Total
committed to using technology to Total Relay to Trip Interrupt Trip During Trip Other Reclose Misoperations
reduce system costs. Real world Events K-Factor Misoperations Fault Than Fault
resource limitations limit how much 20 Not CalculatedSee Right Above 400 0% 0% Not Determined 30% 5% 35%
and how fast optimized solutions can
be implemented. Analysis of system 7 301-400 0% 0% 14% 0% 14%
events can help prioritize updating of 49 201-300 2% 4% 4% 12% 22%
equipment to save the most money. 13 101-200 0% 0% 15% 31% 46%
The operation of a protective
relay can be measured by its security 5 51-100 0% 0% 0% 0% 46%
against false operation, its dependabil- 705 4.8-51* 1% 0.6% 2.5% 2.1% 0.14% 6.4%
ity to operate for faults in its zone of
* Not reported voltage in Working Group I17 Report
pc_v2 10/19/04 10:24 AM Page 38


• Relay Component Failure setting that accounts for paralleling sources or other changed
• Relay Design Hole system conditions. Between these two solutions, setting-caused
• Accessory Component Failure or coordination-caused false trips would be reduced by over
• Setting or Coordination Failure 50%.
• Human Caused
• Force Majoure ACCESSORY COMPONENT FAILURE
Looking at a Pareto categorization of causes for misoper- Even though accessories are not a direct part of a relay,
ations, we can see where a small effort can produce maximum they are certainly part of a relay system. The 12 misoperations
improvement. There is some ambiguity in classifying these have only three root causes, all of which can be addressed with
events. Is a false trip caused by a wiring error, an accessory low-cost modern technology.
component failure, a human caused failure, or a relay design Six false trips were caused by copper pilot wires being
hole? Is a false trip caused by a shorted pilot wire relay, an shorted, which caused a false trip on an external fault. Three of
accessory component failure, or a relay design hole? In both these shorts appeared to be long-term failures while three of
cases they are categorized as an accessory component failure, them were caused when a fault on a nearby line brought down a
although it could be claimed that the relay is at fault for not phase conductor and damaged the pilot wires. This demon-
detecting the problem. In the broadest sense, all failures could strates the importance of both long-term monitoring of commu-
be classified as human caused in that piling enough systems on nication channels as well as high-speed supervision of trips with
top of each other can prevent any error. For the sake of simplic- a loss of channel signal.
ity, the first error cause is used as the classification. False trips caused by bad wiring were the second leading
By category, the misoperations are rated and ranked by cause, with five occurrences. Three of these were differential
percentage of failures: relays that had been miswired on initial installation. Using a
• Setting or Coordination Failure: 18 instances (27%) relay capable of displaying phase rotation and steady state oper-
• Accessory Component Failure: 12 instances (18%) ating quantities provides a means of checking secondary CT and
• Human Caused: 12 (8 of these were due to one break-in by VT wiring.
vandals) (18%) One false trip was caused by an electromechanical auxil-
• Relay Design Hole: 9 instances (13.5%) iary relay continually keying a permissive transmitter until an
• Induced Signal/Noise: 5 instances (7.6%) external fault caused a misoperation. Again, the use of a modern
• Force Majoure: 5 instances (7.6%) relay with a channel monitor and timer alarm could prevent this
• Relay Component Failure: 3 instances (4.5%) operation.
• Mystery: 2 instances (3%)
Drilling into the specifics of these categories, it can be HUMAN-CAUSED MISOPERATION
seen where protective design action can prevent misoperations. With eight trips caused by vandals during one break-in, it
is not difficult to determine the primary cause of human-caused
SETTING OR COORDINATION FAILURE misoperations. The station break-in occurred at 6:04 p.m. in an
Line differential relays had the most coordination prob- urban area on Thursday, May 8, 1997. This demonstrates that
lems (five), all due to fuses on taps on the lines. In some physical security may not be sufficient to prevent breaker oper-
instances, a time-overcurrent relay had been added to provide a ation by unauthorized persons. There are operational questions
level of coordination with tapped fuses. This seemed to elimi- about routing all trips through password-controlled devices, but
nate the coordination problem at the cost of fairly long time if cases like this one increase, the incentive to use the secure
delays and increased cost in material and wiring. devices available will increase.
Four false trips, on three different protection schemes, The other human-caused events were either bumping
were caused by system conditions that were not considered relay panels and RTU racks or dropping wires. The bumping
when applying settings. Two false trips, one of them on a 345 demonstrates the need for high seismic contacts. The bump of
kV line, were caused by delayed or repeated tripping of adjacent the RTU rack showed no targets of any kind following the oper-
lines. Twice, frequency relays operated for transient conditions. ation. Routing through a relay with SER would provide a record
While most frequency relays are used with a time delay, of trips and their initiating device. De-bounce timers on inputs
system faults and switching operations can cause transients that can help prevent cascading a bump-caused event.
fool a frequency relay under some conditions. These operations
can cause a greater system problem than a false line trip because RELAY DESIGN HOLE
reclosing is not normally performed, but waits for operator To some extent, any relay false trip can be described as a
intervention. Other setting-caused false trips were from incor- design hole. For this analysis, however, the definition is restrict-
rect echo signals, and in one case, from new settings that were ed to those cases where the relay clearly misoperated even
not installed when a breaker configuration was changed. though it tested as OK. Considering the system exposure, the
None of the setting or coordination misoperations were low number of operations of this type is a credit to the qualifi-
the result of a missed calculation or the wrong curve selection. cation testing performed by relay manufacturers and utilities.
The combination of tapped loads on differentially protected The most common design-caused misoperation (five
lines and conditions not modeled made up the bulk of these occurrences) was distance relays operating on either a PT fail-
cases. Both of these conditions can be addressed with modern ure or a remote fault causing a low voltage on a radialized sys-
relays. Tapped load coordination using the sum of both line end tem. These misoperations were all on electromechanical relays.
currents provides shorter coordinating margins than with a sin- Solid-state and microprocessor relays had no recorded loss-of-
gle end time-overcurrent relay supervising the differential relay. potential operations.
Multiple setting groups can use external inputs to change to a The second greatest single design hole based on these 19
pc_v2 10/19/04 10:24 AM Page 39


months of experience was three cases of electromechanical failure to operate for a fault is of great concern to the relay engi-
transformer differential relays (built by two different manufac- neer. Local backup can limit damage and the spread of tripping,
turers) operating on inrush. Considering that inrush is itself a but loss of service will certainly be greater than if the primary
statistical phenomenon, with only about one out of 20 energiza- relay operates correctly. By the time remote backup takes place,
tions producing maximum inrush, this number can be consid- numerous lines must be cleared. Because fault current is divid-
ered significant. ed between these lines, the delay in clearing is significant.
Microprocessor relays provide improved setting ranges The fault shown in Appendix II that occurred on 5/5/96
and operating principles [2]. Possibly of greater importance, is lasted at least 82 cycles (the fault recorder stopped recording at
the recorded waveform of all trip events to provide analysis that point. Six lines connecting to six different stations cleared
without the time and expense of testing the transformer itself. to remove the fault current. Murphy’s law was strongly evident,
The only other false trip from a relay that tested OK was as it took a dispatcher five tries to find the correct breaker to
a solid-state phase comparison relay that operated for a fault on open in order to restore the system. With minimal fault data
a parallel line. It is possible this is a setting error. The limited available, problems compounded and it took 35 minutes before
information available from a solid-state relay makes a detailed lines tripped on backup could be closed.
diagnosis impossible. Using the same categories and rankings as listed for
misoperations, we can group the failure to trip events as fol-
INDUCED SIGNAL/NOISE lows:
Four of the five instances of induced signal-caused or • Setting or Coordination Failure: 1 instance (7.7%)
noise-caused trips were in communication circuits, not in the • Accessory Component Failure: 10 instances (76.9%)
relays themselves. The use of communications dependent • Human-Caused: 0 instances
schemes for EHV protection makes these events even more sig- • Relay Design Hole: 0 instances
nificant. At 500 kV, one event was with phase comparison and • Induced Signal/Noise: 1 instance (7.7%)
one was with differential. This demonstrates the problems with • Force Majoure: 0 instances
using a communication system subject to noise, with a protec- • Relay Component Failure: 1 instance (7.7%)
tion scheme dependent on accurate communications. Direct or • Mystery: 0 instances
multiplexed fiber systems would be more appropriate for com-
munication-dependent protection schemes. SETTING OR COORDINATION FAILURE
The other instance, by itself, continues to illustrate the The only event caused by a setting problem was a fault
problems with having unrecorded trip paths. A circuit breaker below set pickup in a time-overcurrent relay.
operated during a DC ground search with no relay targets
recorded. With no record of a device operation, any corrective ACCESSORY COMPONENT FAILURE
action is through guesswork only. Accessory component failure, including wiring, was
responsible for ten times more failures to trip than any other
FORCE MAJOURE cause. Clearly attention needs to be paid to the root cause and
These may be the only five instances of false trips that detection of these errors. In analyzing these failures, it is fortu-
defy a reasonable corrective action. The industrial explosion nate that for the ten failures to trip there were only three basic
that caused the three line relays to operate broke windows a half causes.
mile from the facility and the substation was across the street. A For five of the failures, the circuit breaker failed to trip.
reasonable case could be made that the trips were correct. The failures were caused by both trip coil failures and mechan-
The other two events were water gaining access to trans- ical or electrical failure of the breaker itself. While trip coil
former mounted relays. Proper maintenance of gaskets could monitoring has been available for years in solid-state and micro-
help, but the relays were mounted where they needed to be. processor relays, it is not used in all applications. More com-
plete monitoring of actual circuit breaker interruptions has only
RELAY COMPONENT FAILURE recently been available in microprocessor relays. Even using a
At almost the bottom of the list of false trips was relay simple breaker history report (see Figure 1) can provide infor-
component failure. Of concern is that in all three instances the mation on the “health” of a circuit breaker before a failure to trip
component failure in an electromechanical or solid-state relay clears at least a system bus. The utility presented did not use
system was found as a result of a system fault. The implication breaker failure protection at sub-transmission levels, and remote
is that more quietly failed components are waiting for a nearby clearing was typically delayed.
event to trigger a false trip. In this case again, a relay with self- Four failures were caused by shorted or miswired pilot
checking diagnostics does not need to wait for a false operation wires. Combined with the false trips detailed above, this is the
to determine that there is a problem. greatest single failure mode. This amounted to almost a 1%
chance of failure for all events. Pilot wire relays were on only a
MYSTERY little more than 10% of feeders. With roughly one chance in ten
The last two false trips could almost certainly fit into one of a misoperation or failure to operate, the pilot wire failure rate
of the categories listed above, but there is no data to draw from. seems unacceptably high. Monitoring of differential communi-
No targets, test failures, or event records indicate the guilty cations or replacement of copper wire with optical fiber should
devices. Again, the importance of recording relays is indicated. be a very high priority for any system.
One failure to trip was caused by the CT wires being
FAILURE TO OPERATE (SEE APPENDIX II) reversed for a directional overcurrent relay. A three-phase fault
had current of 5200 amps at 34.5 kV for 15 seconds. From a
In terms of equipment damage, if not system damage, the power standpoint alone, this is 1.29 Megavar hours. This was a
pc_v2 10/19/04 10:24 AM Page 40

Fig.1. Breaker History Report
new electromechanical installation at the time. A microproces- At sub-transmission levels (34.5 kV) there is a large
sor relay would have shown reversed CTs with a glance at the amount of data provided by the utility. This can be analyzed sta-
meter report. tistically to examine the performance of these relays and the
system impact this speed has.
INDUCED SIGNAL/NOISE Figure 2 shows the sub-transmission TOC clearing times
One instance is a data point, not a trend. Arcing noise versus fault duty. While this has a “shotgun” look, and clearing
blocked a power line carrier signal in a POTT scheme at 230 kV. times that do not change a lot with increased fault duty, there are
This resulted in tripping by a backup system in 24 cycles. If it other ways to look at the data that is even worse. By removing
were a trend, switching to a blocking scheme or changing a reclose tries (which have faster clearing times due to preloading
Directional Comparison Unblocking (DCUB) scheme using relay current) the clearing times can be seen to be even slower
Frequency Shift Keying (FSK) would be reasonable. as illustrated in Figure 3. For comparison, a moderately inverse
TOC curve is included to show the relationship that current and
RELAY COMPONENT FAILURE time could reasonably be expected to have for a protected sys-
With only one component-failure caused failure to trip tem.
out of over 1400 events, the lack of instances is of more interest Of course a single TOC curve is only applicable to a sin-
than the one event. gle line. By stacking curves for coordination, it is expected that
clearing times actually increase as a fault is closer to the source.
SECURITY/RELIABILITY TRADE-OFFS This tradeoff of higher speed for higher currents, and lower
speed as a fault moves closer to the source is what results in the
The one event of component failure in a 55-year-old scatter plots shown. The negative results of this are clearly seen
relay shows the exemplary reliability of relay systems in the in the minimum clearing speeds as a function of fault duty. At
past decades. This should make us think about the traditional 7000 Amps of fault duty there is a minimum clearing time of
use of two different relay systems for increased reliability. The five cycles. At 10,000 Amps that has gone up to 7–8 cycles, and
data provided by this utility indicates that almost all failure to at 20,000 Amps it has increased further to 10 cycles.
trips are caused by connected wires or circuit breaker problems,
not relay construction or design. This indicates that adding a
second, dissimilar, relay system produces virtually no increase
in protection system reliability. On the other hand, because a
second, dissimilar, relay system increases the probability of set-
ting errors, the probability of a false trip roughly doubles with
the added relay. Where two relays are desired for maintenance
or testing purposes, this data shows that having similar wiring
and settings will provide the least security degradation.
PROTECTION SPEED
After security and reliability, the next measure of a pro-
tection system’s performance is speed. The data source for this
paper indicates EHV protection times of 2–4 cycles for most
faults. There are numerous papers discussing protection speeds
at EHV levels so this paper will not go into further detail [3],
[4], [5].
Fig.2. Time Current Points Chart
pc_v2 10/19/04 10:24 AM Page 41


er speed tripping.
In addition, the likelihood of damage to nearby lines was
limited by faster tripping. Specifically in TOC protected lines,
there was a 31% probability of a failed first reclose, compared
to a 25% probability for differentially protected lines of the
same class. Also, in 5% of the faults, a second fault would be
caused by the first. These were all on TOC protected lines, with
the second fault occurring at least six cycles after the first with
a typical time of 30–60 cycles after the first fault started.
An example of how fault times add up is a storm event
that occurred on January 6, 1997. Numerous lines experienced
multiple temporary faults, followed by a high-speed reclose
operation. One typical example is the 34.5 kV line identified as
the V-P3 line. It experienced 15 faults over an eight-hour peri-
od, with an average magnitude of over 7000 Amps and the dura-
tion 11 of each fault of 45 cycles. This is 11 seconds of fault
duty. The transformer feeding this line survived the day, but it
Fig.3. TOC Without Reclosing Tries Chart could not have felt good.
This type of event can be improved in two ways. Fault-
clearing time can be reduced with relayto- relay communica-
By comparison, some lines of the utility are protected
tions. This average clearing time of 45 cycles could reasonably
with differential relays. For all of their lack in security and
be reduced to six cycles. Microprocessor relays can also
dependability, it cannot be argued that differential relays are
improve the system operation by counting operations in a given
fast.
time and limiting relatively long term events to a reasonable
Figure 4 shows the clearing times for faults on a differ-
level. If a slack span is causing numerous trips in a storm, it
entially protected line. Note the difference in scales from the
might be a better idea to trip off the line until the span can be
TOC curves.
repaired than to continue hammering it with fault after fault.
CONCLUSIONS
Based on the data from this utility we can make some
general conclusions. It is recognized that, because the data is
from only one utility, care should be taken in its unquestioned
application, however the large number of events sampled
increases confidence.
1. False trips outnumber failure to trips by a factor of
about five to one. Protection quality improvements should be
focused with this ratio in mind.
2. Failures to trip are only rarely caused by relay failures
or design flaws. Doubling overall protection scheme complexi-
ty can decrease security without improving reliability unless
steps are taken to minimize the possibility of setting or accesso-
ry problems.
Fig.4. Time Current Differential Chart 3. Pilot communications are the number one cause of
both security and reliability problems. Improving the quality of
communications will have a direct benefit to protection system
One of the advantages of microprocessor relays is that quality.
high-speed protection does not require expensive relays. The 4. Measured protection speed at all voltage levels should
same relay that is used for directional time-overcurrent protec- be examined for suitability and reasonability to limit equipment
tion can be used in a high-speed protection scheme with the sim- and system damage.
ple addition of low-cost communications [3].
IEEE standard C57.12 for Liquid-Immersed Distribution, REFERENCES
Power, and Regulating Transformers [4] states that “…the dura- [1] IEEE Power System Relay Committee Working
tion of the short-circuit current as defined in 7.1.4 is limited to Group I17 Report, Transmission Relay System Performance
2 s, unless otherwise specified by the user.” It goes on to say that Comparison.
this time includes all reclosing operations. It is well understood [2] Armando Guzman, Stan Zocholl, Gabriel Benmouyal,
that fault current and time will damage transformers, with each Hector J. Altuve, “Performance Analysis of Traditional and
event taking a piece of life out of the unit. One of the elements Improved Transformer Differential Protective Relays,”
of a high-quality protection system is that equipment damage be Technical Paper, 2000.
limited, as much as reasonable, given the overall economics of [3] James R. Fairman, Karl Zimmerman, Jeff W.
the protection. Gregory, James K. Niemira, “International Drive Distribution
In addition to reduced equipment damage, the data Automation and Protection,” Proceedings of the 27th Annual
reveals that reclosing was somewhat more successful with high- Western Protective Relay Conference, Spokane, WA, October,
pc_v2 10/19/04 10:24 AM Page 42


2000. Communications Processors, RTUs, and PLCs as Substation
[4] IEEE Standard, C57.12 for Liquid-Immersed Automation Controllers,” Technical Paper, October 17, 2000.
Distribution, Power, and Regulating Transformers.
[5] Edmund O. Schweitzer III, Ken Behrendt, Tony Lee, Roy Moxley has a B.S. in Electrical Engineering from the
“Digital Communications for Power System Protection,” University of Colorado. He joined Schweitzer Engineering
Proceedings of the 25th Annual Western Protective Relay Laboratories in 2000 as Market Manager for Transmission
Conference, Spokane, WA, October, 1998. System Products. Prior to that he was with General Electric
[6] Gabriel Benmoyal, Jeff Roberts, “Superimposed Company as a Relay Application Engineer, Transmission and
Quantities: Their True Nature and Application in Relays,” Distribution (T&D) Field Application Engineer, and T&D
Proceedings of the 30th Annual Western Protective Relay Account Manager. He is a registered Professional Engineer in
Conference, Spokane, WA, October, 2003.
the State of Pennsylvania.
[7] David J. Dolezilek, “Choosing Between
APPENDIX I: FALSE OPERATIONS

RELAY COMPONENT FAILURE
4/8/96 Reclosing relay with shorted diode, closed in three times, loss of air pressure in circuit breaker caused trip times to increase until backup relay (on 230 kV bank) cleared fault on 34.5 kV feeder.
3/15/96 Staged fault caused adjacent 500 kV line to trip by “finding” a faulty component that removed restraint and caused operation on reverse fault. This sent a direct transfer trip to the other end.
6/29/97 230 kV line tripped due to leaking capacitor in electromechanical distance relay.
RELAY DESIGN HOLE
Two electromechanical distance relays operated for remote bus fault “the relay contacts have a history of drifting closed when the line voltage goes dead.” They didn’t cause outage. The line
1/30/96 was already dead.
8/11/96 Solid-state phase comparison relay tripped for a fault on parallel line. Relays were tested with no problems found.
9/11/96 Electromechanical distance relays tripped on pt failure, line did not trip.
9/23/96 Electromechanical transformer differential misoperated during inrush. Relay tested OK.
E/M DCB scheme misoperated at one end of line due to fault detector operating for external fault and forward looking distance relay “drifting” closed on low voltage (two occurrences on sepa-
9/25/96 rate lines for same fault).
ACCESSORY COMPONENT FAILURE
1/27/96 9:41 Electromechanical pilot wire differential false trip on bad pilot.
1/27/96 9:48 Electromechanical pilot wire differential false trip on bad pilot.
8/1/96 E/M POTT scheme false tripped on external fault due to e/m aux failure causing transmitter to stay keyed on.
8/1/96 Solid state Bus differential tripped on external fault due to a ground return wire not installed during addition of new equipment to station.
9/18/96 Three transformer banks tripped due to false transfer trip during test of breaker failure relays. Blocking switches were mislabeled on newly installed equipment.
11/20/96 Directional overcurrent relay opened while switching a capacitor due to a control wiring problem.
1/6/97 Fault on adjacent line damaged pilot wires, causing Electromechanical pilot wire differential relays to trip three lines.
5/6/97 Electromechanical pilot wire differential tripped on external fault. Apparently shorted pilot.
6/24/97 Transformer false tripped on first load because CT wired backwards.
7/8/97 Same transformer tripped again due to one phase wired incorrectly.
SETTING OR COORDINATION FAILURE
1/16/96 Electromechanical pilot wire differential operated on fuse cleared fault. Electromechanical pilot wire differential can’t coordinate with fuse, cleared faults.
3/15/96 500 kV staged fault caused a echo tripping permissive echo that eventually caused a false trip on that line. Line tripped again on second staged fault test on adjacent line.
3/18/96 Over-frequency relay tripped on transient caused by line tripping. Relay operated correctly, given its settings, but incorrectly given its application.
3/25/96 Relay operated for a repeated fault on an adjacent 345 kV line. This was a “correct” incorrect operation. Could be described as a coordination failure.
4/5/96 Transfer trip inadvertently sent during disconnect switching 230 kV line.
4/5/96 Electromechanical pilot wire differential tripped after fuse cleared fault—lack of coordination.
5/17/96 Electromechanical pilot wire differential tripped after fuse cleared fault—lack of coordination.
7/4/96 Electromechanical pilot wire differential false tripped due to circulating current when transformers were paralleled.
9/19/96 4.8 kV bus tripped on backup due to slow trip of downstream fault (coordination failure).
12/12/96 Overcurrent relay on transformer tripped on backup when a fault on a feeder did not clear—Coordination error.
1/16/97 Underfrequency relays tripped on the transient when a breaker tripped on low SF6 pressure. Setting error in my opinion.
2/27/97 EM TOC Relay tripped on circulating current when bus tie closed for routine work.
3/21/97 Electromechanical pilot wire differential over-tripped on fault cleared by fuse tapped on line.
4/4/97 Electromechanical pilot wire differential tripped due to circulating current when lines paralleled.
5/19/97 EM directional overcurrent tripped when line was paralleled.
5/23/97 Electromechanical pilot wire differential over-tripped on fault cleared by fuse tapped on line.
pc_v2 10/19/04 10:24 AM Page 43

SETTING OR COORDINATION FAILURE (C0NTINUED)

7/3/97 Transformer relay false tripped on new energization because new settings had not been applied.
INDUCED SIGNAL/NOISE
3/15/96 Staged fault at a 500 kV line caused false trips due to noise induced into phase comparison relay at same station, which sent a transfer trip to other end.
7/23/96 Breaker tripped due to a spike in the DC circuit during a DC ground search. No relay targets were reported.
10/16/96 Electromechanical pilot wire differential relay misoperated due to external 230 kV fault sending “noise spike” into pilot wires, which tripped one end of 34.5 kV line.
12/17/96 Fault on nearby line created a voltage spike causing a pilot wire relay to operate (line did not have drainage reactor).
8/23/97 500 kV false-trip due to microwave noise causing current differential relay to operate.
MYSTERY
3/18/96 230 kV line tripped for fault on reverse line. No targets found on any relay.
8/27/96 230 kV bus tripped during transfer of station service. No targets, no cause found.
HUMAN CAUSED
4/25/96 500 kV line tripped on transfer trip accidentally sent during maintenance.
11/4/96 Electromechanical pilot wire differential false tripped when “a construction crew was drilling on the adjacent relay panel when the relay was jarred closed.”
12/31/96 Transformer tripped when RTU was bumped, causing it to operate. No relay targets (shows advantage of using relay trip contacts for operation).
3/8/97 False trip of transformer due to wiring being dropped into a pool of water during work on transformer pressure relay.
5/8/97 Vandals broke into substation. Tripped eight breakers. No relay targets. Another reason to use relays to operate breakers. Break-in at 6:04 p.m.
FORCE MAJOURE
2/20/96 Water leaked into Buchholz relay.
11/11/96 “Concussion from a large explosion at X caused the relay contact to close” EM directional overcurrent relay (three lines).
1/13/97 False trip due to rain water leaking into the pressure relay on a LTC.
APPENDIX II: FAILURE TO OPERATE

SETTING OR COORDINATION FAILURE
2/16/96 Electromechanical TOC relay didn’t operate for fault 1000 Amp. Cleared other end after 63 cycles. Fault self-cleared at 125 cycles.
ACCESSORY COMPONENT FAILURE
1/29/96 6:36, CB failed to trip (reported as relay failure to trip)
Electromechanical pilot wire differential at 34.5 kV failed to operate due to miswired ground lead which allowed an induced voltage to counteract the tripping voltage. This caused 6 line trips
5/5/96 followed by five reclosing & trips. Dispatcher could not determine where the fault was and closed in repeatedly to test lines.
Breaker failed during trip for line fault (E/M POTT). Failure caused a bus fault to be detected. Breakers on the bus were blocked from tripping due to a large pump being started causing
12/28/96 breaker failure of all incoming 230kV feeds.
12/28/96 After clearing of the breaker fault, station was attempted to re-energize. Fault was re-initiated and same problems happened again.
1/6/97 E/M directional OC relay failed to trip due to CT’s being reversed. Backup tripping cleared five incoming lines at 34.5 kV. Fault took approximately 15 seconds to clear.
1/6/97 Failure to trip Electromechanical pilot wire differential due to shorted pilot wires. Line cleared on time-overcurrent backup.
Failure to trip Electromechanical pilot wire differential due to shorted pilot wires. Line cleared on time-overcurrent backup. This was a repeat event two minutes following a successful reclose. It
1/6/97 could be argued that if the pilot wire relay had tripped the damage would have been limited and reclose would have held… maybe.
1/6/97 Failure to trip EM TOC due to bad breaker. Breaker would not open until all current flow was interrupted elsewhere. Cleared two other lines.
1/6/97 Pilot wire shorted caused failure to trip of Electromechanical pilot wire differential. Two lines were cleared in backup.
8/5/97 Failure to trip due to burnt trip coil (EM relays) two lines cleared on backup.
INDUCED SIGNAL/NOISE
Failure to trip of 230 kV E/M POTT primary protection scheme for the line caused by excessive noise from an arcing conductor swamped out the power line carrier receiver. Line tripped on
1/6/97 backup after 24 cycles.
RELAY COMPONENT FAILURE
Electromechanical pilot wire differential failed to trip due to bad “rectox unit” in 55-year-old relay. After failure relays were replaced by similar vintage relays. Six lines tripped as a result of
5/24/96 failure to trip.
pc_v2 10/19/04 10:24 AM Page 44


pc_v2 10/19/04 10:24 AM Page 45

HOW CERTAIN FEATURES OF PROTECTIVE DEVICES

HELP TO PROTECT WORKERS
By Jeanne K. Ziobro, Eaton Corporation
There are more articles, papers, and seminars available relay coordination studies can then be performed.
on the subject of arc flashes, as the hazard is so potentially Today protective relays devices, such as the FP5000 from
harmful not only to individuals or the electrical equipment Eaton Electrical/Cutler Hammer, are equipped with four inde-
being worked on, but also to the companies involved. pendent protection settings groups. This allows a user to config-
This paper seeks to inform the reader of the potential haz- ure each group of protection settings for a different application.
ards involved, some of the standards written and an application Additionally, the FP5000 is equipped with eight optically isolat-
of a protective device and a control scheme to reduce the arc ed inputs that may be configured for custom logic applications.
flash hazard. Connecting the source voltage to the appropriate terminal on the
What is arc flash? According to the NFPA, the developer terminal block activates these contact inputs. Using these two
of the NFPA70E, the standard for Electrical Safety in the features together, a simple control system can be created to
Workplace, 2004 Edition, an arc flash can be defined as “an change the active setting group (from standard to arc fault)
electric current that passes through air when insulation or isola- when personnel are planning to work on energized switch-
tion between electrified conductors is no longer sufficient to boards, panel boards, or motor control centers. The use of a
withstand the applied voltage. The flash is immediate, but the keyed selector switch and other accessories will be required for
results of these incidents can cause sever injury including this application.
burns”. Lets walk through an example scenario: Company A has
How is this different from a bolted fault? A bolted fault just finished performing a short circuit analysis. They use this
current results from phase conductors becoming solidly con- information in preparation for setting their protective devices
nected together, causing large amounts of short-circuit current (in this case, the FP5000) for fault coordination of feeder and
to flow through two or more phase conductors. Arc fault current main breakers. Using the IEEE 1584 spreadsheets, they also
results from phase conductors making less than solid contact, determine the minimum and maximum potential arc fault cur-
causing a necessary arc to sustain the current flow through the rents. Once all calculations are made and protective settings are
loose connection. Usually, the air is the conductor during an arc selected based on the calculated bolted fault and arc-fault cur-
fault. Depending on the arc impedance, the arc current may be rents, the logic is then ready to be selected.
as low as several amperes or as high as 20-40 percent of a three- Using the Contact Inputs settings, logically map Settings
phase, bolted short-circuit current. Since typical arc fault cur- Group 1 (Standard Settings /Normal Mode) to Contact Input 8
rents are less than the currents involved in a bolted fault, the and Settings Group 2 (Arc Fault Settings/Maintenance Mode) to
conventionally configured overcurrent protective functions for a Contact Input 7 (as an example). See Table 1.
bolted fault may not be effective for an arc fault. Externally wire Contact Input 7 to position “b” of a
Although the cause of the short-circuit current normally keyed selector switch labeled Arc Fault Settings/ Maintenance
burns away, the resultant arc fault is sustained and conducts as Mode and Contact 8 to position “a” of a keyed selector switch
much energy as is available depending on the impedance of the labeled Standard Settings/Normal Mode. See Figure 1.
arc. As a result, a delay or failure to clear an arc fault could Therefore, when the selector switch is in position “a”,
result in damage to personnel or equipment. Reported resultant Standard Settings will be in place, when the selector switch is in
damage ranges from structural building damage to personnel position “b”, Arc Fault Settings will be in place. Thus, when
eardrum rupture, blindness or even death. It is these results that personnel approach the switchgear for maintenance, they will
have caused the heightened awareness of arc faults. (with authorization, of course) switch the selector switch to
How do we protect our personnel and equipment? Well, Maintenance Mode, enabling the more sensitive settings based
the only true safe way would be to de-energize equipment on arc fault.
before approaching it for opening or maintenance. Since this is In summary, there is a growing concern for arc flash haz-
typically not practical, companies must find other ways. Using ards and numerous Industry standards developed to address
some standards created and new technology in protective these hazards and help protect unsuspecting personnel. Not only
devices there are ways to help reduce the risk of an arc flash
hazard through the use of protective devices and a selector
switch. TABLE 1. LOGIC SETTINGS FOR FP5000 CONTACT INPUTS 7 & 8
Following the guidelines of the IEEE 1584, a company Contact Input 7 Contact Input 8
should perform a short circuit analysis, as would typically be Settings Group 2 Settings Group 1
done to properly set protective relays. In addition to calculating Setting (Arc Fault Settings/ Maintenance Mode) (Standard Settings /Normal Mode)
the bolted fault current, the potential arc fault current should be active active
calculated (using the established IEEE 1584 formula for calcu- Selector Switch
lating/estimating the arc fault current). Taking these values, Position “b” Position “a”
Position
pc_v2 10/19/04 10:24 AM Page 46


are the standards committees involved but many manufacturers
of electrical equipment are also doing their part to protect the
unsuspecting personnel. This paper explained one method to
enhance your protective relay system coordination and mainte-
nance procedures to protect personnel when performing mainte-
nance on energized switchboards, panel boards and/or motor
control centers.
Jeanne K. Ziobro is a product manager for power quality pro-

tective relay devices at Eaton Corporation. She can be reached
at jeanneziobro@eaton.com.
Fig.1. Selector switch wiring to Contact Inputs 7 & 8 of the FP5000
(Standard Settings /Normal Mode active)
pc_v2 10/19/04 10:24 AM Page 47

CHOOSING THE RIGHT SOLID STATE RELAY
By Lyle W. Strode, President, HBControls, Inc.

For decades, mercury and electromechanical relays have The potential for the use of solid state relays continues to
successfully been in use, and, although health and environmen- expand as companies begin to see the competitive advantage
tal concerns have led many countries to discourage the use of they provide. Due to fewer service calls from their own cus-
mercury, both these relay types may still be appropriate choices tomers, many users of solid state switching devices have been
in many applications. However, where high-speed switching is able to substantially increase the warranties on their products
required, precise control is an issue, electrical noise must be over what is possible when using electromechanical or mercu-
minimized, or relay life is the weak link in the life of a product, ry relays.
solid state relays are far more appropriate. Their initial cost With the knowledge that solid state relays, when proper-
may be higher, but they keep working years longer and, when ly selected and installed, can operate indefinitely, why would a
properly applied, will pay for themselves many times over. manufacturer who needs to control an electrical circuit delay
Many companies are converting to solid state relays in conversion to solid state relays? In many cases it is because the
their products, and long life is often the main reason. Users are process can be confusing to engineers. The interpretation of
able to enhance the value of their products by offering longer load ratings, R-theta requirements, heat generated, ambient tem-
warranties, due to the extended life of the solid state relay. peratures, maximum baseplate temperatures, and other factors
Solid state relays have been proven in many industries. requires complicated calculations. A simple mistake in these
In the plastics industry, for example, solid state relays control calculations can lead to the premature failure of a solid state
multiple heaters in extrusion and injection molding machines. relay, often due to overheating or power anomalies that could
Solid state relays facilitate precisely controlled heating in the have been prevented.
semiconductor, medical, and food-processing industries as well. First, the engineer must thoroughly examine the charac-
Even some vending machines use solid state relays to maintain ter of the load to be switched and the conditions in which the
temperature, keeping foods hot or cold. solid state relay will operate. Typically, the load being con-
But temperature control is only one way of putting solid trolled is a single- or three-phase circuit with a range as high as
state relays to work. The once mysterious little black boxes 660 volts AC. The control signal from the controller can be
now control pumps, motors, solenoids, power contactors, and either AC or DC voltage, and models are available for virtually
elevator door controls. One manufacturer uses solid state relays all types of controllers. The typical solid state relay assembly is
to control the lights, fans, and electrical outlets in a laboratory designed to perform reliably in an ambient temperature of up to
fume hood. Another recently installed more than 500 solid state 40°C, however systems can be designed for higher ambient tem-
relays at a telecommunications switching station. peratures.
Rated for 50 amps, this compact solid state relay assembly can handle a variety of applica- This typical solid state relay assembly is rated for 90-amp loads. The appropriately sized
tions. Standard features in this product line include finger-safe covers, zero voltage switch- anodized aluminum heat sink ensures reliability. A load-suppression diode is standard on
ing, and DIN rail mounts. many models. Relays like this one have been proven in food processing, semiconductor
manufacturing, plastic molding, and many other applications for the control of heaters,
pumps, motors, solenoids, power contactors, lighting, and other equipment.
pc_v2 10/19/04 10:24 AM Page 48

Solid state relay assemblies can be modified

with a Linear Phase Angle Controller, which
offers true linear AC output phase angle control
and accepts a variety of inputs.
A great variety of solid state relay (SSR) assemblies are available, with a choice of mounting options. There are many others in addition Another variation on the solid state relay assem-
to the basic models shown here. It is no wonder that plant engineers often seek the advice of SSR experts before deciding on the best bly is a Solid State Relay Monitor, which fits
assembly for the job at hand. snugly on a relay and signals the loss of AC or
DC line current, an open load, or a short circuit.
The unit shows a green LED to indicate normal
operation.
Once knowledgeable engineers know the details, they quately cooled at 40°C or less ambient temperature.
can specify an appropriate, cost-effective relay assembly tai- Relay manufacturers provide maximum temperature
lored to the job at hand. The primary components of these information for their relays. The 25A relay option will perform
assemblies should consist of properly sized and matched relays properly while switching a 20A load with a maximum baseplate
and heat sinks. Since heat is a solid state relay’s primary enemy, temperature of 105°C. The 50A relay will perform properly
the choice of the right heat sink and relay combination is criti- while switching a 20A load with a maximum baseplate temper-
cal to a successful relay assembly. Solid state relay monitors, ature of 112°C.
current transducers, and fault-sensing devices are also available The manufacturer also indicates that both the 25A and the
for consideration. 50A relays generate 20 watts of heat. Knowing the maximum
A cursory look at an example application illustrates typi- allowable temperature, the ambient operating environment, and
cal calculations that must be applied when choosing the proper the watts generated by the relays, the engineer should now be
solid state relay assembly. In this example, the relay is being able to calculate their ability to conduct heat. However, the
used to control a single heater that has a current load of 16 kilo- relay manufacturer states that when operating at maximum rat-
watts, in a three-phase delta configuration on a 480V line with ing, there is a voltage drop across the relay. This requires an
all three legs being switched. The controller in the application adjustment in the value of the relay’s ability to conduct heat.
provides a 24VAC output with a 20MADC minimum current It is important to remember that the smaller this value,
required. The relay will be housed in an enclosure along with the larger the heat sink required. Also, three-phase lines tend to
other electronic equipment. be electronically noisy, and because heaters and their wiring are
Since most electronic equipment begins to experience inductive the energy stored in them could show up across the
degradation at temperatures above 40°C, we have assigned an relay output as a high-voltage transient, in some cases exceed-
ambient operating environment of 40°C MAX. ing 1200V. Therefore, the relay must have a PIV (peak inverse
In order to select the available relay choices for this voltage) rating of at least 1200V, and it also must be protected
application, the first step is to determine the current flow in each by the use of metal oxide varistors (MOVs) or other transient-
line to be switched. Once the load current and line voltage are suppression devices.
known, with the additional knowledge that the line is a three- Given the above information, there are two options for
phase delta configuration, one is able to calculate the amperage satisfying the switching needs of this application.
to be switched at 19.25A per leg. Option 1 uses a three-phase relay mounted on a single
Given this information, there appear to be two available heat sink. The relay chosen draws 53MA at 24VDC and is com-
options for the relay: a 25A or a 50A model could be considered. patible with the controller being used in this application. MOVs
Both will be able to switch the load, provided they are ade- or other transient-suppression devices are recommended in the
pc_v2 10/19/04 10:24 AM Page 49


resultant assembly.
Option 2 uses three individual single-phase relay assem-
blies functioning as a three-phase switch when the inputs are
wired in parallel. Attention must be paid to the total current
requirements of the three relays used in this example. Since the
controller used requires 20MA at the output, if the relays used
do not draw at least 20MA total, a load resistor of the appropri-
ate ohmic value must be wired across the relay input in order to
draw the minimum current required by the controller. As with
Option 1, MOVs or other transient-suppression devices are rec-
ommended in the resultant assembly.
The final choice between the two methods illustrated can
be made by the designer.
One provider of solid state relay assemblies,
HBControls, makes selection easy by designing relay assem-
blies that perform reliably for a wide variety of specified current
loads while operating in specified ambient operating environ-
ments.
Lyle W. Strode, President of HBControls, Inc., has 40 The HBControls website (www.hbcontrols.com) includes an interactive solid state relay selec-
years of experience in the electronics controls industry, special- tion guide.
izing in the applications of electromechanical and solid state
switching devices.
a leader in the design, manufacture, and marketing of DIN Rail
An interactive solid state relay selection guide is avail-
Mounted Solid State Relays and Custom Assemblies. Additional
able on the HBControls website: www.hbcontrols.com.
products from HBControls include Dual Relays, Solid State
HBControls, Inc., based in Fall River, Massachusetts, is
Relay Monitors, and Current Transducers.
pc_v2 10/19/04 10:24 AM Page 50


pc_v2 10/19/04 10:25 AM Page 51

ASPECTS OF OVERCURRENT PROTECTION FOR

FEEDERS AND MOTORS
By E. 0. Schweitzer III, Fellow IEEE, S. E. Zocholl, Fellow IEEE, Schweitzer Engineering Laboratories, Inc.
ABSTRACT applications may appear diverse and unrelated. However,
This paper discusses the coordination of the characteris- microprocessor relay technology has advanced to where it is not
tics and the backup and redundancy possible with microproces- only feasible, but it is of distinct economic advantage to consid-
sor relays. The paper reviews the application of negative- er all these characteristics collectively as attributes of a univer-
sequence overcurrent characteristics for unbalanced protection sal overcurrent relay. Furthermore, the issues of backup and
in motors and also covers the rules for coordinating negative- redundancy are addressed since a single processor easily
sequence characteristics to provide sensitive phase-to-phase accommodates the computational burden of two complete and
protection in feeders. It also covers reset characteristics and the independent relays.
requirements for stator and rotor thermal protection of induction Figure 1 shows the block diagram for a dual universal
motors. overcurrent relay having one opto-isolated input, two output
contacts, and a set of three-phase current inputs for each of the
INTRODUCTION independent relays X and Y. In the relay setting procedure, the
user is prompted for relay X or Y and then for the application.
A comprehensive list of non-directional overcurrrent The relay then presents a group of elements to be set. Such an
relays would include thermal overload, inverse-time, definite arrangement focuses attention, not only on the individual char-
time, and instantaneous relays. The list could be further classi- acteristics, but also on the elements that make up each applica-
fied by operating quantities including individual phase, residual, tion. This universal relay concept is used here to discuss the
and negative-sequence current. Taken collectively and depend- commonality, the differences, and the coordination of the ele-
ing on the characteristic shape, pickup and time range, and ments required for feeder, motor, and breaker failure protection.
dynamics, these relays span the applications for motor, feeder, The paper goes on to discuss the rules for the coordination of
and breaker failure protection. negative-sequence overcurrent characteristics for sensitive
Because of the past necessity for using either discrete or phase-to-phase fault protection in feeders, as well as for unbal-
specialized system relays, overcurrent characteristics for these anced current protection of induction motors.
Fig.1. Relay Block Diagram

pc_v2 10/19/04 10:25 AM Page 52

MEASURANDS AND THE DIGITAL FILTERING

PROCESS
The input for a digital relay is obtained by sampling sine
wave currents at discrete time intervals. A fixed number of
instantaneous samples per cycle are converted to digital quanti-
ties by the AID converter and stored in memory for processing.
Digital Filtering is the simple process of combining successive
samples to obtain the quantities which represent the phasor
components of the input. For example, in Figure 1 six currents
are sampled 16 times per cycle. Each successive sample is mul-
tiplied by the coefficients of a stored cosine wave and combined
to obtain a phasor component. The cosine filter has the follow-
ing properties ideal for protective relaying [1] :
1. Bandpass response about the system frequency.
2. DC and ramp rejection to guarantee decaying exponen-
tials are filtered out.
3. Harmonic attenuation or rejection to limit effects of non-
linearities.
4. Reasonable bandwidth for fast response.
5. Good transient behavior.
6. Simple to design, build, and manufacture. Fig.2. Extremely Inverse Characteristic Compared with Min. Melting and
The present samples and the samples occurring a quarter- Maximum Clearing Time of a SOE Fuse
cycle earlier form the real and imaginary components of the cur-
rent phasors. The measurands calculated from the phasor com- Figure 2 shows the close coordination of an extremely inverse
ponents each cycle are: induction characteristic with that of a high-voltage fuse.
Phase Current Ia, Ib, Ic The straight line I2t log-log plot of a fuse minimum melt-
Residual Current I0 ing time is often visualized as the basic time-current character-
Negative-Sequence Current I2 istic. However, a definite time must be added to form the max-
Positive-Sequence Current I1 imum clearing time characteristic of the fuse. This illustrates the
fundamental concept that whenever fixed clearing time is added
APPLICATION GROUPS AND ELEMENTS to a straight line log-log plot the result is a curve. For this rea-
All the elements of the universal relay are organized son, the most viable shape for a time-current characteristic for
under the applications of feeder, motor, or breaker. The feeder coordination purposes is the curve formed when a definite time
application group provides three separate sets of Instantaneous, is added to the straight line of a 108-108 plot.
Definite-Time, and Inverse-Time elements. The first set Appendix I shows that, were it not for the use of satura-
responds to Ia, Ib, and Ic and provides conventional phase fault tion, the induction characteristic would be the straight line log-
protection. The second set responds to 3I2 and provides for sen- log characteristic of a fuse. However, the viable curve is formed
sitive phase-to-phase fault, and ground backup protection. The by deliberately saturating the electromagnet at a specific multi-
third set responds to residual current 3I0, and provides conven- ple of pickup current to introduce a definite time component as
tional ground fault protection. discussed in [2]. Therefore, adding a constant definite time term
The motor application group shares the phase, negative- to Eq. 12 of Appendix I forms the induction characteristic equa-
sequence, and ground instantaneous and definite-time elements tion. Consequentl,. all the characteristics can be accommodated
to protect the motor for winding faults or faults in the connect- in the dual overcurrent relay using the following equations:
ing leads. However, an added element, responsive to both posi-
tive-sequence and negative-sequence current, provides thermal For 0 < M < 1
protection for overload, locked rotor, or unbalanced current con-
ditions. t = TD ⎛ Α ⎞ (1)
⎝M2 - 1⎠
COORDINATION OF INVERSE TIME-CURRENT
For M > 1
CHARACTERISTICS
Coordination practice is ultimately determined by the t = TD ⎛ Α + B⎞ (2)
⎝M2 - 1 ⎠
type of grounding used in distribution systems. Notably, in
Europe the practice is to operate high impedance grounded or where: t is the trip time in seconds
ungrounded 3-wire distribution systems. Since there are no sin- M is multiples of pickup current
gle-phase laterals protected by fuses, coordination is obtained TD is the time-dial setting (1 through
using definite-time characteristics. Conversely, in North 15)
America the practice is to operate grounded 4-wire distribution
systems with loads served by single-phase laterals protected by where the constants A, B, and the exponent p determine the
fuses. Consequently, coordination is obtained using inverse shape of the characteristic. The constants A and B can be cho-
time-current characteristics suitable for fuse coordination. sen to accurately emulate the extremely and the very inverse
pc_v2 10/19/04 10:25 AM Page 53


induction time-current characteristics with the exponent p of 2. nation time and not by load.
An accurate emulation of the moderately inverse characteristic 4. Retain the time-dial or time-delay settings but multiply
is obtained by using an exponent of 0.02 with specific values for ing the "equivalent" phase-overcurrent pickup setting by
A and B. The shapes of the induction characteristics are defined .√3. The result is the negative-sequence overcurrent ele-
by the ratios of A to B and the exponents listed below: ment pickup in terms of 3I2 current.
5. Set the next upstream negative-sequence element to
Curve A/B p coordinate with the first downstream negative-sequence
Moderately Inverse 0.46 0.02 element (and the third with the second and so on).
Inverse 33.05 2.0
Very Inverse 40.29 2.0 MOTOR PROTECTION
Extremely Inverse 161 2.0 The characteristics of the motor application are shown in
Figure 3 with the starting current of an induction motor. In the
motor application, definite-time and instantaneous elements
COORDINATING NEGATLVE-SEQUENCE provide protection for faults in the motor leads and internal
OVERCURRENT ELEMENTS faults in the motor itself. A definite-time setting of about 6
It is well understood that ground overcurrent elements, cycles allows the pickup to be set to 1.2 to 1.5 times locked rotor
operated by the residual current 3I0, do not respond to balanced current to avoid tripping on the initial Xd" inrush current (shown
load and can be set to operate faster and more sensitively for the magnified). The instantaneous can then be set at twice the
most frequent fault type, the phase-to-ground fault. Similarly, locked rotor current for fast clearing of high fault currents.
negative-sequence overcurrent elements, responding to 3I2, do An inverse-time phase overcurrent element and a sepa-
not respond to balanced load and can be set to operate faster and rate negative-sequence overcurrent element could be applied to
more sensitively for the second most frequent fault type, the prevent the overheating caused by a locked rotor or an unbal-
phase-to-phase fault. ance current condition. However, neither of these elements can
Negative-sequence elements are useful in clearing faults account for thermal history or track the excursions of tempera-
on the secondary of a delta-wye. Also, the settings realized for ture. Instead, an element is used which accounts for the I2r heat-
a bus negative-sequence element may be as sensitive as the ing effect of both positive and negative-sequence current. The
feeder phase-overcurrent elements for effective backup for element is a thermal model, defined by motor nameplate and
feeder phase-to-phase faults. The feeder relay sensitivity for thermal limit data, that estimates motor temperature. The tem-
phase-to-phase faults may also be improved. perature is then compared to thermal limit trip and alarm thresh-
olds to prevent overheating for the abnormal conditions of over-
TABLE 1: FAULTS ON A RADIAL LINE load, locked rotor, too frequent or prolonged starts, and unbal-
Fault | 3I2/Ip | anced current.
AG 1
BC √3
BCG <√3
TABLE 2: SEC. FAULTS ON DELTA-WYE TRANSFORMER

Fault | 3I2/Ip |
AG √3
BC 1.5
The negative-sequence elements differ only in their oper-

ating quantity and are easily coordinated with phase and ground
relays. Elneweihi et al. [3] , have devised a simple method for
setting negative-sequence elements. The method is based on the
observation that the greatest ratio between the negative-
sequence current 3I2 and the phase current Ip is √3 as illustrated
in Tables I and II. The simple method of setting negative-
sequence overcurrent elements is as follows: Fig. 3. Motor Characteristics Plotted with Motor Starting Current
1. Start with the most downstream negative-sequence ele-
ment (e.g., feeder relay).
2. Identify the phase-overcurrent device (relay, fuse, etc.)
downstream from the negative-sequence element that is DEFINING THE THERMAL MODEL
of the greatest concern for coordination. This is usually The I2r heat source and two trip thresholds can be dis-
the phase-overcurrent device with the longest clearing cerned from a motor characteristic of torque, current, and rotor
time. resistance versus slip shown in Figure 4. The plot shows the
3. Consider the negative-sequence element as an "equiva- characteristic of the induction motor to draw excessively high
lent" phase-overcurrent element. Derive the setting for current until the peak torque develops near full speed. Also, the
this element as any phase element would be done. The skin effect of the slip frequency causes the rotor resistance to
setting only differs in that it is only governed by coordi- exhibit a high locked rotor value labeled Rl which decreases to
pc_v2 10/19/04 10:25 AM Page 54


R R ⎛R
⎯r+|S=0 = 1 ⎯r- |S=0 = 2 ⎯1⎞ − 1 = 5 (7)
R0 R0 ⎝R0⎠
These factors are the coefficients of the positive and neg-

ative currents of the heat source in the thermal model.
STATES OF THE THERMAL MODEL

Because of its torque characteristic, the motor must oper-
ate in either a high-current starting state or be driven to a low-
current running state by the peak torque occurring at about 2.5
per unit current. The thermal model protects the motor in either
state by using the trip threshold and heating factors indicated by
the current magnitude. The two states of the thermal model are
shown in Figure 5. The thermal model is actually a difference
equation executed by the microprocessor. However, it can be
Fig. 4. Current, Torque. and Rotor Resistance of an Induction Motor Versus Speed represented by the electrical analog circuit shown in Figure 5. In
this analogy, the heat source is represented by a current genera-
a low running value at rated slip labeled R0. tor, the temperature is represented by voltage, and thermal
Using a typical starting current of six times the rated cur- resistance and capacitance are represented by electrical resist-
rent and a locked rotor resistance Rl of three times value of R0, ance and capacitance. The parameters of the thermal model are
the I2r heating is estimated at 62 x 3 or 108 times normal. defined as follows:
Consequently, an extreme temperature must be tolerated for a R1 = Locked rotor electrical resistance (per unit ohms)
limited time to start the motor. Where an emergency I2t thresh- R0 = Running rotor electrical resistance also rated slip (per
old is specified by the locked rotor limit during a start, a thresh- unit ohms)
old for the normal running condition is specified by the service IL = Locked rotor current in per unit of full load current
factor. Therefore, the thermal model requires a trip threshold Ta. = Locked rotor time with motor initia1ly at ambient
when starting, indicated by the locked rotor thermal limit, and a T0 = Locked rotor time with motor initia1ly at operating
trip threshold when running, indicated by the service factor. temperature
The slip dependent beating effect of positive- and nega-
tive-sequence currents is derived as follows. The positive- The starting state is shown in Figure 5a and is declared
sequence rotor resistance plotted in Figure 4 is calculated using whenever the current exceeds 2.5 per unit of the rated full load
current, torque, and slip in the following equation: current and uses the threshold and heating factors derived for
the locked rotor case. Thermal resistance is not shown because
QM the start calculation assumes adiabatic heating. The running
Rr = ⎯⎯ S (3) state, shown in Figure 5b, is declared when the current falls
I2 below 2.5 per unit current and uses the heating factors derived
for the running condition. In this state the trip threshold "cools"
and can be represented as a linear function of slip. The positive-
exponentially from a locked rotor threshold to the appropriate
sequence resistance Rr+ is a function of the slip S:
threshold for the running condition using the motor thermal
time constant. This emulates the motor temperature which cools
Rr+ = (R1 -R0)S + R0 (4)
to the steady-state running condition.
In the model, the thermal limit IL2Ta .represents the
The negative-sequence resistance Rr- is obtained when S
locked rotor hot spot limit temperature and IL2(Ta-T0) represents
is replaced with the negative-sequence slip (2-S):
the operating temperature with full load current. The locked
rotor time, Ta is not usually specified, but may be calculated by
Rr- = (R1 -R0)(2 - S) + R0 (5)
using a hot spot temperature of six times the operating temper-
ature in the following relation:
Factors expressing the relative heating effect of positive-
and negative-sequence current are obtained by dividing Eqs. 4 IL2Ta T
and 5 by the running resistance, R0. Consequently, for the ⎯⎯⎯⎯⎯ =6 ∴ ⎯a = 1.2 (8)
IL (Ta - T0)
2
T0
locked rotor case, and where Rl is typically three times R0, the
heating effect for both positive- and negative-sequence current There are two reasons for using the rotor model in the
is three times that caused by the nomal running current. running state. The first is that, despite a difference in thresholds,
R R R1 it is an industry practice to publish the overload and locked rotor
⎯r+|S=1 = ⎯r- |S=1 = ⎯ =3 (6) thermal limits as one continuous curve as illustrated in Figure 3.
R0 R0 R0
The second is that the rotor model accounts for the heating of
For the running case, the positive-sequence heating fac- both the positive-sequence and the negative-sequence current.
tor returns to one and the negative-sequence heating factor As a final refinement, assigning standard values of 3 and 1.2 to
increases to 5: the ratios R1/R0 and Ta/T0 respectively, allows the model
parameters to be determined from five fundamental settings:
pc_v2 10/19/04 10:25 AM Page 55

CONCLUSION
1. An analytic equation for induction-type inverse time-
current characteristics has been derived. The integral equation
also defines reset characteristic and the dynamics which guaran-
tee close coordination with induction relays under all conditions
of varying current.
2. The equations of a motor thermal element have been
derived. The element exists as two-state thermal which accounts
for the slip-dependent heating of positive- and negative-
sequence current. The model is defined by motor nameplate and
thermal-limit data and provides protection during the abnormal
conditions of overload, locked rotor, and unbalanced current.
3. The overcurrent elements applied for feeder, motor,
and breaker protection consist of sets of thermal, inverse-time,
definite-time, and instantaneous elements. These elements,
grouped by application, are collectively accommodated as
attributes of a universal overcurrent relay.
4. The issues of backup and redundancy have been
addressed by a dual-relay implementation.
5. Negative-sequence elements, as well as traditional
phase and ground elements, are obtained from three-phase cur-
Fig.5. States of the Thermal Model rent measurands. Negative-sequence elements with induction-
type characteristics coordinate directly with phase and ground
FLA Rated fu1l-load motor current in secondary amps elements and can be set independent of balanced load to provide
LRA Rated locked rotor current in secondary amps sensitive phase-to-phase fault coverage.
LRT Thermal-limit time at rated locked rotor current
TO Time dial to trip temperature in per unit of LRT APPENDIX I
SF Motor rated service factor THE TIME-CURRENT EQUATION
An equation for the inverse time-current characteristic
DUAL APPLICATIONS can be derived from the following basic differential equation for
Out of the many possible applications, two shown in input dependent time delay as it applies to an induction relay:
Figures 6a and 6b illustrate the versatility of the dual universal dθ
overcurrent relay. In Figure 6a, both the relays X and Y are set τs(Μ2 − 1) = Kd ⎯ (9)
for feeder application to protect a delta-wye transformer bank. dt
Relay X provides phase and negative-sequence overcurrent pro- Where:
tection on the high side (delta) that also see through the bank to M is the ratio I/Ip
the low side. The ground overcurrent elements provide sensitive I is the input current in amperes
protection for the high side but cannot see through the delta. Ip is the pickup current in amperes
However, relay Y provides the ground protection for the low τs is the spring torque
side. In Figure 6b, relay X is set as a feeder where the phase and Kd is the damping factor due to the drag magnet
negative-sequence overcurrent elements provide protection for θ is the angular displacement and dθ/dt is the
high phase-to-phase faults using a higher ratio ct. Relay Y is set angular velocity
for motor application using a low ratio ct to protect the small
motor . The small moment of inertia of the disc is neglected and
the spring torque is represented by a constant because the effect
of its gradient is compensated by an increase in torque caused
by the shape of the disc. Integrating Eq. 9 gives:
T τs
θ = ⌠ ⎯0
(M2 - 1)dt (10)
⌡0 Kd
Dividing both sides of Eq .10 by 8 gives the dynamic

equation:
T τs
0
T 1 0
⌠ ⎯⎯ (M2 - 1)dt = ⌠ ⎯⎯ dt = 1 (11)
⌡0 Kdθ ⌡0 t(I)
where t(I) is the time-current characteristic and the con-

stant A equals Kdθ/τs:
Fig.6. Applications of a Dual Universal Overcurrent Relay

pc_v2 10/19/04 10:25 AM Page 56


(Kdθ/τs) A current relays under all conditions of varying current such as
t(I) = ⎯⎯⎯⎯ = ⎯⎯⎯ (12) decreasing fault resistance and remote terminal clearing.
(M2 - 1) (M2 - 1)
RESET CHARACTERISTIC REFERENCES

Where the time integral of any element can be reset in [1] E. 0. Schweitzer and Daqing Hou, "Filtering For
one cycle, an optional reset characteristic is available when Protective Relays", 19th Annual Western Protective Relay
required for close coordination with existing induction relays. Conference, Spokane, WA, October 1992.
Eq. 12 defines the induction characteristic for currents [2] G. Benmouyal
below, as well as for currents above, the pickup current. , and S. E. Zocholl, "Testing Dynamic Characteristics of
If an induction disc has an initial displacement from its Overcurrent Relays", .PEA Relay Committee Fall Meeting,
reset position when the applied current is reduced to zero, the Hershey, PA, September 21-22, 1993.
disc will be driven in a negative direction toward the reset posi- [3] A. F. Elneweihi, E. 0. Schweitzer,. and M. W. Feltis,
tion. This is represented in Eq. 12 by setting M = 0, which pro- "Negative-Sequence Overcurrent Element Application and
duces a negative number indicating the reset time and the rota- Coordination in Distribution Protection",. IEEE Transactions on
tion of the disc in the direction toward reset. With this substitu- Power Delivery. Volume 8, Number 3, June 1993, pp 915-923.
tion, Eq. 12 gives the reset time tr: [4] S. E. Zocholl, E. 0. Schweitzer III, and A. Aliaga-
Zegarra, "Thermal Protection of Induction Motors Enhanced by
Kdθ Interactive Electrical and Thermal Models", IEEE Transactions
| tr | = ⎯⎯ (13) on Power Apparatus and Systems, Volume PAS-I03, Number 7,
τs
July 1984, pp 1749-1755.
and the reset characteristic for any value of M between zero and [5] B. N. Gafford, W. C. Duesterheoft Jr., and C. C.
one is: Mosher III, "Heating of Induction Motors on Unbalanced
tr Voltages", AIEE Transactions, June 1959, pp 282-287.
t = ⎯⎯ (14) [6] J. H. Dymond, "Stall Time, Acceleration Time,
M2-1 Frequency of Starting: The Myths and the Facts", IEEE
Transactions on Industrial Applications, Volume 29, Number 1,
The dynamic equation, Eq. 11, and the characteristic January/February 1993.
equation, Eq. 12, are important since they specify how an [7] M. Shan Griffith, "A Penetrating Gaze at One Open
inverse time-current characteristic must be implemented in Phase: Analyzing the Polyphase Induction Motor Dilemma, "
order to guarantee coordination with existing inverse-time over- IEEE Transactions on Industry Applications, Volume IA-13,
pc_v2 10/19/04 10:25 AM Page 57

THERMALLY PROTECTED MOV
By Ferraz Shawmut
INTRODUCTION Thermally Protected MOV. The philosophy to circuit protection
MOVs used in SPDs can fail explosively when subject- is no damage protection, and the TPMOV offers this.
ed to sustained steady-state power frequency over-voltages.
Traditional MOVs are highly susceptible to damage from sus- FERRAZ SHAWMUT THERMALLY PROTECTED
tained/temporary over-voltage conditions (Fig. 1). MOV(TPMOV)
One example of an over-voltage condition may occur The Ferraz Shawmut Thermally Protected MOV has
when the neutral conductor has been removed, either deliberate- been developed to assist in minimizing the failure characteris-
ly or accidentally, from a split phase or three-phase wye config- tics of Metal-Oxide-Varistors. It is composed of a voltage
uration. Another may be a misapplication of the product in a clamping device, two forms of isolated indication and a discon-
higher than rated voltage system. In both cases the available necting apparatus that monitors the status of the metal-oxide
voltage is greater than the maximum continuous operating volt- disk (Fig. 2).
age (MCOV) specified by the manufacturer of the MOV. During In the event that the disk has been or is approaching
these over-voltage conditions MOV(s) enter a conductive state,
absorb the associated energy, generate a great deal of heat and
eventually rupture, initiating a short circuit condition.
Protection against the destructive consequences of MOV
failure is provided by :
1. Current limiting fuse to reduce the damage in the event
of MOV failure (the fuse by itself only minimizes the
damage, it does not eliminate the condition).
2. Thermal fuse to detect and disconnect MOV from serv-
ice in the event the disk temperature exceeds specified
levels.
3. Filter networks.
4. Some designs use packed sand, epoxy and other materi- Fig. 2.
al around MOV to assist in limiting the MOV failure
from propagating surrounding equipment.
The prominent failure mode for MOVs is the sustained breakdown, it is disconnected from system power. The TPMOV
over-voltage condition. Under this condition the MOV begins to is rated to withstand 40ka of 8/20µs surge current (Fig. 4). The
conduct power frequency (50-60hz) current. Ultimately, the MCOV ratings available are from 150VAC to 550VAC. The
MOV will reach its maximum energy capacity and go into a TPMOV has been designed with two built in isolated indicating
thermal runaway condition. Under normal operating conditions, features. The first is a visual indicator composed of two pins
the MOV absorbs random transient currents (short in duration), that proceed through the top of the TPMOV in the event that the
transforming the energy into heat. The MOV has a finite energy device has disconnected from system power. This indicator
capacity. When this capacity is exceeded, the MOV can no allows the operator to see which device has broken down with-
longer effectively dissipate the heat and ultimately shorts (in the out the shutting off all system power, to determine which device
milli-ohms). For these reasons Ferraz Shawmut developed the in array has been degraded. The second feature is a remote indi-
cator composed of a N/O 12VDC micro-switch. In the event
that the TPMOV has disconnected system power the switch will
change status to closed. This feature lowers costly engineering
time that would be required for traditional MOV products.
TRADITIONAL MOV FAILURE MODE AND TPMOV

SOLUTION
For sustained steady-state over-voltages a typical test
requirement for an MOV is that it should withstand 125% of its
rated voltage for a given duration. However, MOVs may fail to
a short-circuit condition when subjected to sustained steady-
state over-voltages above this level. In service such over-volt-
ages may be caused by abnormal system operating conditions,
Fig. 1. MOV Failure Due to sustained Over-Voltage but a more common situation is that the SPD unit is incorrectly
pc_v2 10/19/04 10:25 AM Page 58


selected or wired and failure occurs after initial switch-on. tor must have a thermal response which coordinates closely
Failure under these conditions is caused by bulk overheating of with the response of the MOV. Fig. 3 shows the behavior of a
the MOV. One approach would be to use MOVs with higher 320V MOV when subjected to a sustained over-voltage of
knee voltages but this gives poorer transient over-voltage 150% (480V). The supply circuit had an available short-circuit
clamping levels. A better approach is to use a thermal protector, current of 41.2 kA at 15.2% power factor. In Fig. 3(a) the cur-
which responds to the temperature of the MOV and disconnects rent wave for an ineffectively thermally protected MOV is
the MOV if this temperature exceeds a certain level. The MOV shown. The MOV conducts during the part of the a.c. cycle
is required to absorb the energy from a specified test current where the supply voltage exceeds the knee voltage of the MOV,
surge, typically an 8/20µs wave with a 10kA peak, and the ther- and the amplitude of the current pulses depends on the supply
mal protector must be able to withstand this without opening. impedance and the MOV resistance. The temperature of the
The TPMOV accomplishes these tasks by withstanding MOV increases and this leads to failure of the MOV after 0.176
40ka of surge current and offers very fast disconnect response in seconds. Fig. 1(b) shows the corresponding MOV voltage.
the event that the MOV disk has been degraded. Ferraz Fig. 3(c) shows the response of the Ferraz Shawmut
Shawmut has conducted extensive testing on a variety of SPDs TPMOV (Thermally Protected MOV). The current pulses are
and has witnessed extensive damage due to ineffective electri- higher than in Fig. 3(a) because of normal variation in MOV
cal protection. In our research, there are many MOV-based tran- characteristics. In this case, the TPMOV operates after 0.045
sient suppressors that utilize both short-circuit and thermal fus- seconds and isolates the MOV with zero arcing.
ing. Short-circuit fusing alone does not provide complete pro- If the available short-circuit current is reduced to a level
tection. Many of these designs rely on the thermal sensing to where the on resistance limits MOV conduction, the amplitude
disconnect under limited current conditions (less than 10 amps) of the current pulses fall and the operating times increase.
and rely on the short-circuit fusing to disconnect under high Nevertheless, the TPMOV still operates prior to failure.
fault current conditions.Both of the mentioned conditions are a
function of the energy capacity of the MOV-based suppressor. TPMOV SURGE CAPACITY
Under limited current conditions, the MOV shows limited signs The TPMOV has been designed with the ability to dis-
of physical damage but can create enormous amounts of heat connect in the event of thermal degradation and the ability to
over a specific amount of time (enough time to transfer heat to withstand 40ka of IEEE spec 8/20µs surge current waveform
melt the thermal element). The vast majority of these thermal (Fig. 4). The TPMOV is designed with a specific waveform
fuses don't work under high fault current applications because capacity. In order to achieve effective thermal protection the
of the location of the thermal sense. The thermal fuse or thermal capacity of surge needed to be established with minimal I^2t at
leg is not located close enough to the MOV disk in order to 40ka. The waveform utilized to meet the 40ka peak current has
sense the actual temperature. With high available fault currents a 6us front time and a 17µs tail time (Fig. 4). The waveform tol-
the traditional MOV cannot effectively transfer heat to the ther- erances are specified within IEEE C62 for surge protection
mal sense prior to thermal runaway. The Ferraz Shawmut products. If the TPMOV is subjected to surges in excess of the
TPMOV has a thermal sensing means that reacts to the actual specified energy capacity, the thermal sense will disconnect. As
MOV disk status(Fig. 3). For this reason, the TPMOV can dis- is with all MOVs, if the incoming surges are greater than it’s
connect from 6mA to 100ka of available fault current (the avail- ability to dissipate the heat energy, the MOV will be degraded.
able short circuit current does not depict the TPMOV’s perform- Realistically when in use, the TPMOV will very seldomly be
ance. subjected to surges greater than 10ka.
Complete electrical protection for SPDs must include
short circuit protection and thermal sensing. The thermal protec- INDICATOR RESPONSE:
The TPMOV has two forms of built in indication. It is
critical to disconnect very quickly in the event the MOV disk
has been degraded, it is also very important to detect quickly
when the transient protector has been disconnected from sensi-
tive electronics. The TPMOV can disconnect as quickly as 20ms
(1.5 cycles) and indication is not more than 2ms from the time
Front Time 6 us
Tail Time 17 us
Total I^2t 17.5kA^2s
Circuit Parameters
L = 1.7uH
R = .25ohms
V = 19kV
F = 25uF
Fig.3a, b, c. Response of Ferraz Shawmut TPMOV to Sustained Over-Voltage condition Fig.4.

pc_v2 10/19/04 10:25 AM Page 59


the thermal leg is removed from MOV
disk (Figure shown).
TPMOV TRANSIENT VOLTAGE SUPPRESSION:

superior clamping ability to sup-
press incoming high over-voltage tran-
sients. The TPMOVs clamping ability is
exceptional because users can reduce the
series components used to protect
against thermal runaway. The TPMOV
has built-in thermal protection with very
little area between the MOV disk and
mounting location. This approach
reduces the effect of induced voltages
added to the clamping ability during
high energy/high frequency transients.
Fugure 6 represents the clamping abioity
of a high amplitude incoming transient.
The top line (current curve) represents
the shunted transient current. The second
line (clamped voltage curve) represents
a highly damaging transient voltage and
the third line (transient voltage surge)
represents how the TPMOV suppresses
the incoming over-voltage to levels that
will not damage sensitive electronic
devices or equipment.
CONCLUSIONS
SPDs need to be protected
Before Activation against the 60Hz fault currents which
After Activation
follows an MOV failure. The Ferraz
Shawmut TPMOV offers a safety level
Fig. 5. to engineers utilizing an MOV based
suppressor that has not been available in
the past.
Fig. 6.
pc_v2 10/19/04 10:25 AM Page 60


pc_v2 10/19/04 10:25 AM Page 61

RIO GRANDE ELECTRIC MONITORS REMOTE

ENERGY ASSETS OVER SATELLITE
By Anthony Tisot, Power Measurement
Rio Grande Electric Cooperative (RGEC) in
Brackettville, Texas, provides electric service to over 10,000
metering points across Texas and New Mexico. Traditionally,
the company’s commitment to providing a reliable source of
power across 27,000 square miles of barren terrain has present-
ed some interesting challenges - just to check and service
remote substations, maintenance crews routinely faced driving
times of up to seven hours each way. So, when RGEC’s
Technical Services department considered upgrades to the co-
op’s electrical metering equipment, the question of communica-
tions was paramount.
Here was an opportunity to monitor the status of the
entire distribution network, while reducing the amount of time
technicians must spend on the road — but how could they com-
municate with remote substations where telephone coverage
was sparse, and cellular networks practically nonexistent? As
part of a planned upgrade to a full ION® enterprise energy man-
agement system from Power Measurement, RGEC identified
satellite communications as the answer. Combining reasonable
costs with unlimited flexibility, satellite stood out as the most Old Metering Cabinet: One of the old metering cabinets to be replaced.
cost-effective way to support reliable communications between
the energy-management software installed at the headquarters, each location. The revenue meter serves as a gateway, collect-
and the network of intelligent energy meters destined for the ing energy data from the feeder meters, and passing it by high-
remote substations. speed Ethernet link to the satellite hardware, and on to a PC
server at the RGEC headquarters equipped with the ION
THE SATELLITE SOLUTION Enterprise™ energy management software.
According to Mike Wade, technical services manager for To help integrate the satellite communications, RGEC
RGEC, cost was a key factor in favor of a satellite link. worked with a local provider of industrial satellite installations.
Although RGEC had selected meters with an available modem “They had a lot of experience installing satellite radios for oil
option, connecting the remote substations via dial-up telephone and gas drilling businesses, and because this type of industrial
link would have required the company to install many miles of application was fairly similar in many respects, there were no
new telephone lines and poles. "We realized that the high cost surprises,” explained Wade. “As Rio Grande’s service area is a
of running a phone line to each of these remote substations - typically hot, desert-like environment, we also decided to build
plus the ongoing fees for using each line per minute - would not cooling hardware into the enclosures; this was a fairly straight-
present an affordable solution," said Wade, "and because cell forward process too.”
phones aren't supported in most of these locations either, a satel- The end result is an enterprise energy management sys-
lite link was the clear choice. With satellite, the cost for each tem that provides real-time power monitoring and control capa-
substation is limited to a satellite system for about $2,500, and bility across the entire distribution network. The system offers
a connection rate of approximately $200 per month. That price 24-hour access to real-time and logged system information for
gets us a full-time Ethernet connection by satellite between the each substation “and the connection is fast,” said Wade,
meters in the field and the energy-management server at the “because it uses Ethernet between the meters and the satellite
head office.” connection, and between the satellite and the master software
station at the head office. This enables a true ‘SCADA’ opera-
INTEGRATED ENERGY MANAGEMENT tion with real-time monitoring of energy and power-quality con-
RGEC worked with Power Measurement to pre-assemble ditions.”
and install new electrical panels for each of its 18 substations.
Each substation or metering point was equipped with a “master” REMOTE NOTIFICATION
ION 8500™ or ION 7600™ revenue-accurate energy meter to The new energy management system also includes an
monitor total power in the substation, and each circuit exiting Internet–enabled alarm feature that instantly notifies key per-
the substation was equipped with an ION 7350™ feeder meter. sonnel of potential areas of concern. Using the system’s unique
A satellite modem and VSAT satellite dish was then installed at
pc_v2 10/19/04 10:25 AM Page 62


helps RGEC engineers to continually
review and improve the company’s
engineering processes, and update its
system distribution model.
INCREASED RELIABILITY
Through increased efficiency,
reduced response time, and improved
awareness, Wade confirms that the
new energy management system has
already helped to identify and correct
potential threats to reliability at sever-
al remote substations. “Recently, we
received an inrush of alarms from our
Cienega substation, indicating multi-
ple transients on one phase.
Inspection of the substation revealed a
damaged bypass arrestor and dam-
aged bushing on the regulator,”
explained Wade. “In another case,
alarms from our Altuda substation
warned of low voltage readings, so
our area office manager dispatched a
RGEC New Substation Panel (with insets): A new substation panel combines an ION 7600(tm) master meter, and two ION lineman to the substation, who then
7350(tm) feeder monitors with satellite radio equipment. identified and corrected a problem
with a regulator control. By automati-
MeterM@il® e-mail messaging feature, personnel can now cally monitoring all of our substations
receive alarm notification for all over/under voltage situations, 24 hours a day, we can ensure that our technicians spend less
sags or swells, transients, or unusually high temperatures with- time on the road, and more time where they’re needed most.”
in a meter enclosure. Also, the system sends an e-mail alarm if With a new substation scheduled for completion in mid
the kVA exceeds 80% of the rated kVA of the power trans- 2004, RGEC has already arranged to bring it online with anoth-
former. er master meter and four feeder monitors, and like the other
Each “master” meter also hosts its own onboard web locations, the new station will be equipped with a VSAT satel-
page, making detailed power system information accessible to lite Internet connection for 24-hour power monitoring and con-
authorized personnel anywhere, through a standard web brows- trol. “Although this satellite-based application represents a fair-
er. At RGEC, Technical Services, as well as Engineering and ly new communications strategy for a rural utility, it’s proven to
Operations staff regularly use this WebMeter® web–enabled be a very useful one,” confirmed Wade. “As we continue to
feature to check on conditions at a specific substation (and save increase efficiency, improve reliability, and reduce operating
themselves a long drive out to the site). costs across our entire distribution network, we can extend these
To monitor consumption across the entire distribution benefits to our members, and for a cooperative like Rio Grande
network, the system automatically records total kWh as interval Electric, that’s the bottom line.”
data logs, and distributes this information as monthly reports to Power Measurement is a leading provider of enterprise energy
TXU — the cooperative’s power supplier, and ERCOT — the
management systems for energy suppliers and consumers
independent system operator. To help RGEC control power
quality and reliability across its distribution network, the system worldwide. For more information, contact Power Measurement
automatically logs line-neutral voltage per phase, amps per at 866-466-7627, or visit the Power Measurement web site at
phase, kW, kVAR, kVA, kWh, power factor, line frequency, www.pwrm.com.
sags, swells, transients, and harmonics. This detailed power data
pc_v2 10/19/04 10:25 AM Page 63

GENERATOR PROTECTION APPLICATION GUIDE

By George Rockefeller, Basler Electric Company
INTRODUCTION The references listed on Page 77 provide more back-

This guide was developed to assist in the selection of ground on this subject. These documents also contain
relays to protect a generator. The purpose of each relay is Bibliographies for further study.
described and related to one or more power system configura-
tions. A large number of relays is available to protect for a wide GROUND FAULT PROTECTION
variety of conditions. These relays protect the generator or The following information and examples cover three
prime mover from damage. They also protect the external impedance levels of grounding: low, medium, and high. A low
power system or the processes it supplies. The basic principles impedance grounded generator refers to a generator that has
offered here apply equally to individual relays and to multi- zero or minimal impedance applied at the Wye neutral point so
function numeric packages. that, during a ground fault at the generator HV terminals,
The engineer must balance the expense of applying a ground current from the generator is approximately equal to 3
particular relay against the consequences of losing a generator. phase fault current. A medium impedance grounded generator
The total loss of a generator may not be catastrophic if it rep- refers to a generator that has substantial impedance applied at
resents a small percentage of the investment in an installation. the wye neutral point so that, during a ground fault, a reduced
However, the impact on service reliability and upset to loads but readily detectable level of ground current, typically on the
supplied must be considered. Damage to and loss of product in order of 100-500A, flows. A high impedance grounded genera-
continuous processes can represent the dominating concern tor refers to a generator with a large grounding impedance so
rather than the generator unit. Accordingly, there is no standard that, during a ground fault, a nearly undetectable level of fault
solution based on the MW rating. However, it is rather expect- current flows, necessitating ground fault monitoring with volt-
ed that a 500kW, 480V, standby reciprocating engine will have age based (e.g., 3rd harmonic voltage monitoring and funda-
less protection than a 400MW base load steam turbine unit. mental frequency neutral voltage shift monitoring) relays. The
One possible common dividing point is that the extra CTs location of the grounding, generator neutral(s) or transformer,
needed for current differential protection are less commonly
seen on generators less than 2MVA, generators rated less than
600V, and generators that are never paralleled to other genera-
tion.
This guide simplifies the process of selecting relays by
describing how to protect against each type of fault or abnormal
condition. Then, suggestions are made for what is considered to
be minimum protection as a baseline. After establishing the
baseline, additional relays, as described in the section on
Extended Protection, may be added.
The subjects covered in this guide are as follows:
• Ground Fault (50/51-G/N, 27/59, 59N, 27-3N, 87N)
• Phase Fault (51, 51V, 87G)
• Backup Remote Fault Detection (51V, 21)
• Reverse Power (32)
• Loss of Field (40)
• Thermal (49)
• Fuse Loss (60)
• Overexcitation and Over/Undervoltage (24, 27/59)
• Inadvertent Energization (50IE, 67)
• Negative Sequence (46, 47)
• Off-Frequency Operation (81O/U)
• Sync Check (25) and Auto Synchronizing (25A)
• Out of Step (78)
• Selective and Sequential Tripping
• Integrated Application Examples
• Application of Multifunction Numerical Relays
• Typical Settings
• Basler Electric Products for Protection Fig.1. Effects of Fault Location Within Generator on Current Level.
pc_v2 10/19/04 10:25 AM Page 64


for ground faults, then the 51N provides the primary protection
for the generator. The advantage of the 87G is that it does not
need to be delayed to coordinate with external protection; how-
ever, delay is required for the 51N. One must be aware of the
effects of transient DC offset induced saturation on CTs during
transformer or load energization with respect to the high speed
operation of 87G relays. Transient DC offset may induce CT
saturation for many cycles (likely not more than 10), which may
cause false operation of an 87G relay. This may be addressed by
not block loading the generator, avoiding sudden energization of
large transformers, providing substantially overrated CTs,
Fig. 2. Ground-fault relaying generator low-impedance grounding. adding a very small time delay to the 87G trip circuit, or setting
the relay fairly insensitively.
also influences the protection approach. The neutral CT should be selected to produce a second-
The location of the ground fault within the generator ary current of at least 5A for a solid generator terminal fault,
winding, as well as the grounding impedance, determines the providing sufficient current for a fault near the generator neu-
level of fault current. Assuming that the generated voltage along tral. For example, if a terminal fault produces 1000A in the gen-
each segment of the winding is uniform, the prefault line- erator neutral, the neutral CT ratio should not exceed 1000/5.
ground voltage level is proportional to the percent of winding For a fault 10% from the neutral and assuming I1 is proportion-
between the fault location and the generator neutral, VFG in Fig. al to percent winding from the neutral, the 51N current will be
1. Assuming an impedance grounded generator where (Z0, 0.5A, with a 1000/5 CT.
Fig. 3 shows multiple generators with the transformer
SOURCE and ZN)>>ZWINDING, the current level is directly propor-
tional to the distance of the point from the generator neutral providing the system grounding. This arrangement applies if the
[Fig. 1(a)], so a fault 10% from neutral produces 10% of the cur- generators will not be operated with the transformer out of serv-
rent that flows for a fault on the generator terminals. While the ice. The scheme will lack ground fault protection before gener-
current level drops towards zero as the neutral is approached, ator breakers are closed. The transformer could serve as a step-
the insulation stress also drops, tending to reduce the probabili- up as well as a grounding transformer function. An overcurrent
ty of a fault near the neutral. If a generator grounding imped- relay 51N or a differential relay 87G provides the protection for
ance is low relative to the generator winding impedance or the each generator. The transformer should produce a ground cur-
system ground impedance is low, the fault current decay will be rent of at least 50% of generator rated current to provide about
non-linear. For I1 in Fig. 1, lower fault voltage is offset by lower 95% or more winding coverage.
generator winding resistance. An example is shown in Fig. 1(b). Fig. 4 shows a unit-connected arrangement (generator
The generator differential relay (87G) may be sensitive and step-up transformer directly connected with no low-side
enough to detect winding ground faults with low-impedance breaker), using highresistance grounding. The grounding resis-
grounding per Fig. 2. This would be the case if a solid generator and voltage relays are connected to the secondary of a distri-
tor-terminal fault produces approximately 100% of rated cur- bution transformer. The resistance is normally selected so that
rent. The minimum pickup setting of the differential relays (e.g., the reflected primary resistance is approximately equal to one-
Basler BE1-CDS220 or BE1-87G, Table 2) should be adjusted third of the single phase line-ground capacitive reactance of the
to sense faults on as much of the winding as possible. However, generator, bus, and step-up transformer. This will limit fault cur-
settings below 10% of full load current (e.g., 0.4A for 4A full rent to 5-10A primary. Sufficient resistor damping prevents
load current) carry increased risk of misoperation due to tran- ratcheting up of the sound-phase voltages in the presence of an
sient CT saturation during external faults or during step-up intermittent ground. The low current level minimizes the possi-
transformer energization. Lower pickup settings are recom-
mended only with high-quality CTs (e.g., C400) and a good CT
match (e.g., identical accuracy class and equal burden).
If 87G relaying is provided per Fig. 2, relay 51N (e.g.,
Basler relays per Table 2) backs up the 87G, as well as external
relays. If an 87G is not provided or is not sufficiently sensitive
Fig. 3. System grounded externally with multiple generators. Fig. 4. Unit-connected Case with high-resistance grounding.
pc_v2 10/19/04 10:25 AM Page 65


enabled. The 59P relay should be set for about 90% of rated
voltage. An “a” contact of the unit breaker can be used instead
of the 59P relay to supervise 27-3N tripping. Blocking the 27-
3N until some level of forward power exists also has been done.
However, use of the 59P relay allows the 27-3N to provide pro-
tection prior to synchronization (i.e., putting the unit on line),
once the field has been applied.
Fig. 5. Neutral shift during ground fault on high impedance grounded system. In order to provide 100% stator winding coverage, the
undervoltage (27-3N) and overvoltage (59N) settings should
overlap. For example, if a generator-terminal fault produces
bility of sufficient iron damage to require re-stacking. Because 240V, 60 Hz across the neutral voltage relay (59N), a 1V pick-
of the low current level, the 87G relay will not operate for sin- up setting (a fairly sensitive setting) would allow all but the last
glephase ground faults. (1/240)*100 = 0.416% of the winding to be covered by the over-
Protection in Fig. 4 consists of a 59N overvoltage relay voltage function. If 20V third harmonic is developed across the
and a 27-3N third-harmonic undervoltage relay (e.g., Basler relay prior to a fault, a 1V thirdharmonic drop-out setting would
relays per Table 2). As shown in Fig. 5, a ground fault at the provide dropout for a fault up to (1/20)*100= 5% from the neu-
generator high voltage bushings elevates the sound phase line to tral. Setting the 59N pickup too low or the 27N dropout too low
ground voltages to a nominal 173% of normal line to neutral may result in operation of the ground detection system during
voltages. Also, the neutral to ground voltage will rise to the nor- normal operating conditions. The third harmonic dropout level
mal phase-ground voltage levels. The closer the ground fault is may be hardest to properly set, since its level is dependent on
to the generator neutral, the less the neutral to ground voltage machine design and generator excitation and load levels. It may
will be. One method to sense this neutral shift is with the 59N be advisable to measure third harmonic voltages at the genera-
relay (Fig. 4) monitoring the generator neutral. The 59N will tor neutral during unloaded and loaded conditions prior to
sense and protect the generator for ground faults over about selecting a setting for the 27-3N dropout. In some generators,
95% of the generator winding. The selected 59N (Basler relays the third harmonic at the neutral may become almost unmeasur-
per Table 2) relay should be selected so as to not respond to third ably low during low excitation and low load levels, requiring
harmonic voltage produced during normal operation. The 59N blocking the 27-3N tripping mode with a supervising 32 under-
will not operate for faults near the generator neutral because of power element when the generator is running unloaded.
the reduced neutral shift during this type of fault. There is also some level of third harmonic voltage pres-
Faults near the generator neutral may be sensed with the ent at the generator high voltage terminals. A somewhat pre-
27-3N. When high impedance grounding is in use, a detectable dictable ratio of (V3RD-GEN.HV.TERM)/(V3RD-GEN.NEUTRAL) will
level of third harmonic voltage will usually exist at the genera- exist under all load conditions, though this ratio may change if
tor neutral, typically 1-5% of generator line to neutral funda- loading can induce changes in third harmonic voltages. A
mental voltage. The level of third harmonic is dependent on ground fault at the generator neutral will change this ratio, and
generator design and may be very low in some generators (a 2/3 this ratio change is another means to detect a generator ground
pitch machine will experience a notably reduced third harmon- fault. Two difficulties with this method are: problems with
ic voltage). The level of harmonic voltage will likely decrease developing means to accurately sense low third harmonic volt-
at lower excitation levels and lower load levels. During ground ages at the generator high voltage terminals in the presence of
faults near the generator neutral, the third harmonic voltage in large fundamental frequency voltages, and problems with deal-
the generator neutral is shorted to ground, causing the 27-3N to ing with the changes in third harmonic ratio under some operat-
drop out (Fig. 6). It is important that the 27-3N have high rejec- ing conditions.
tion of fundamental frequency voltage. If the 59N relay is only used for alarming, the distribu-
The 27-3N performs a valuable monitoring function tion transformer voltage ratio should be selected to limit the sec-
aside from its fault detection function; if the grounding system ondary voltage to the maximum continuous rating of the relay.
is shorted or an open occurs, the 27-3N drops out. The 59P If the relay is used for tripping, the secondary voltage could be
phase overvoltage relay in Fig. 4 supervises the 27-3N relay, so as high as the relay’s ten-second voltage rating. Tripping is rec-
that the 86 lockout relay can be reset when the generator is out ommended to minimize iron damage for a winding fault as well
of service; otherwise, the field could not be applied. Once the as minimizing the possibility of a multi-phase fault.
field is applied and the 59P operates, the 27-3N protection is Where wye-wye voltage transformers (VTs) are connect-
ed to the machine terminals, the secondary VT neutral should
not be grounded in order to avoid operation of 59N for a second-
ary ground fault. Instead, one of the phase leads should be
grounded (i.e., "corner ground"), leaving the neutral to float.
This connection eliminates any voltage across the 59N relay for
a secondary phase-ground fault. If the VT secondary neutral is
grounded, a phase-ground VT secondary fault pulls little cur-
rent, so the secondary fuse sees little current and does not oper-
ate. The fault appears to be a high impedance phase to ground
fault as seen by the generator neutral shift sensing relay (59N),
leading to a generator trip. Alternatively, assume that the VT
corner (e.g., phase A) has been grounded. If phase B or C fault
Fig. 6. Ground fault near generator neutral reduces third-harmonic voltage in generator to ground, the fault will appear as a phase-phase fault, which
neutral, dropping out 27-3N.
pc_v2 10/19/04 10:25 AM Page 66


will pull high secondary currents and will clear the secondary
fuse rapidly and prevent 59N operation. A neutral to ground
fault will tend to operate the 59N, but this is a low likelihood
event. An isolation VT is required if the generator VTs would
otherwise be galvanically connected to a set of neutral-ground-
ed VTs. Three wye VTs should be applied where an iso-phase
bus (phase conductors separately enclosed) is used to protect
against phase-phase faults on the generator terminals.
The 59N relay in Fig. 4 is subject to operation for a
ground fault on the wye side of any power transformer connect-
ed to the generator. This voltage is developed even though the
generator connects to a delta winding because of the trans-
former inter-winding capacitance. This coupling is so small that
its effect can ordinarily be ignored; however, this is not the case
with the 59N application because of the very high grounding
resistance. The 59N overvoltage element time delay allows the
relay to override externalfault clearing.
The Basler BE1-GPS100, BE1-951, BE1-1051, and
BE1-59N relays contain the required neutral overvoltage (59N),
undervoltage (27-3N), and phase overvoltage (59P) units.
Fig. 4 shows a 51GN relay as a second means of detect-
ing a stator ground fault. The use of a 51GN in addition to the
59N and 27-3N is readily justified, since the most likely fault is
a stator ground fault. An undetected stator ground fault would
be catastrophic, eventually resulting in a multiphase fault with
high current flow, which persists until the field flux decays (e.g.,
for 1 to 4s). The CT shown in Fig. 4 could be replaced with a Fig. 8. Medium-level grounding with 87N ground differential protection.
CT in the secondary of the distribution transformer, allowing
use of a CT with a lower voltage rating. However, the 51GN
relay would then be inoperative if the distribution transformer generator terminals.
primary becomes shorted. The CT ratio for the secondary- con- Multiple generators, per Fig. 7, can be highresistance
nected configuration should provide for a relay current about grounded, but the 59N relays will not be selective. A ground
equal to the generator neutral current (i.e., 5:5 CT). In either fault anywhere on the generation bus or on the individual gen-
position, the relay pickup should be above the harmonic current erators will be seen by all 59N relays, and the tendency will be
flow during normal operation. (Typically harmonic current will for all generators to trip. The 51N relay, when connected to a
be less than 1A but the relay may be set lower if the relay filters flux summation CT, will provide selective tripping if at least
harmonic currents and responds only to fundamental currents.) three generators are in service. In this case, the faulted genera-
Assuming a maximum fault current of 8A primary in the neutral tor 51N relay will then see more current than the other 51N
and a relay set to pick up at 1A primary, 88% of the stator wind- relays. The proper 51N will operate before the others because of
ing is covered. As with the 59N relay, the 51GN delay will allow the inverse characteristic of the relays. Use of the flux summa-
it to override clearing of a high-side ground fault. An instanta- tion CT is limited to those cases where the CT window can
neous overcurrent element can also be employed, set at about accommodate the three cables. Fault currents are relatively low,
three times the timeovercurrent element pickup, although it may so care must be exercised in selecting appropriate nominal relay
not coordinate with primary vt fuses that are connected to the current level (e.g., 5A vs. 1A) and CT ratio. For example, with
a 30A fault level and a 50 to 5A CT, a 1A nominal 51N with a
pickup of 0.1A might be used. With two generators, each con-
tributing 10A to a terminal fault in a third generator, the faulted-
generator 51N relay sees 2*10/(50/5) = 2A. Then the relay pro-
tects down to (0.1/2)*100 = 5% from the neutral.
When feeder cables are connected to the generator bus,
the additional capacitance dictates a much lower level of
grounding resistance than achieved with a unit-connected case.
A lower resistance is required to minimize transient overvolt-
ages during an arcing fault.
Ground differential (Fig. 8) is a good method to sense
ground faults on low and medium impedance grounded units. It
would more commonly be seen on generators that have the CTs
required for phase differential relaying. In Fig. 8, the protective
function is labeled 87N, but the Basler BE1-CDS220 or the
BE1-67N is applied. The BE1-CDS220 approach is more appli-
cable to low and medium impedance grounded generators with
Fig. 7. 59N Relay operation within multiple units will not be selective; 51N relays pro- ground faults as low as 50% of phase fault current. The BE1-
vide selective protection if at least three generators are in service.
pc_v2 10/19/04 10:25 AM Page 67


67N approach is more applicable to medium impedance gener-
ators with low ground fault current levels. The BE1-CDS220 is
limited in sensitivity to ground faults in excess of 10% of the
phase CT tap setting, but the use of the auxiliary CT in the BE1-
67N approach allows for amplification of the ground current in
the phase CTs, yielding increased sensitivity. Whichever
approach is used, an effort should be made to select relay set-
tings to trip for faults as low as 10% of maximum ground fault
current levels. During external phase faults, considerable 87N
operating current can occur when there is dissimilar saturation
of the phase CTs due to high AC current or due to transient DC
offset effects, while the generator neutral current still will be
zero, assuming balanced conductor impedances to the fault. One
method to compensate for transient CT saturation is to have suf-
ficient delay in the relay to ride through external high-current
two-phase-ground faults.
PHASE-FAULT PROTECTION
Fig. 9 shows a simple means of detecting phase faults,
but clearing is delayed, since the 51 relay must be delayed to
coordinate with external devices. Since the 51 relay operates for
external faults, it is not generator zone selective. It will operate
for abnormal external operating conditions such as remote faults Fig. 10. Generator fault current decay example for 3 phase and phase-phase faults at
that are not properly cleared by remote breakers. The 51 pickup generator terminals with no regulator boosting or dropout during fault and no pre-fault load.
should be set at about 175% of rated current to override swings
due to a slow-clearing external fault, the starting of a large decay as fast for a phase-phase or a phaseground fault and,
motor, or the re-acceleration current of a group of motors. thereby, allows the 51 relay more time to trip before current
Energization of a transformer may also subject the generator to drops below pickup. Fig. 10 assumes no voltage regulator
higher than rated current flow. boosting, although the excitation system response time is
Fig. 10 shows an example of generator current decay for unlikely to provide significant fault current boosting in the first
a 3 phase fault and a phase-phase fault. For a 3 phase fault, the second of the fault. It also assumes no voltage regulator dropout
fault current decays below the pickup level of the 51 relay in due to loss of excitation power during the fault. If the generator
approximately one second. If the time delay of the 51 can be is loaded prior to the fault, prefault load current and the associ-
selectively set to operate before the current drops to pickup, the ated higher excitation levels will provide the fault with a higher
relay will provide 3 phase fault protection. The current does not level of current than indicated by the Fig. 10 curves. An estimate
of the net fault current of a pre-loaded generator is a superposi-
tion of load current and fault current without pre-loading. For
example, assuming a pre-fault 1pu rated load at 30 degree lag,
at one second the 3 phase fault value would be 2.4 times rated,
rather than 1.75 times rated (1@30°+1.75@90°=2.4@69°).
Under these circumstances, the 51 relay has more time to oper-
ate before current decays below pickup.
Figure 9 shows the CTs on the neutral side of the gener-
ator. This location allows the relay to sense internal generator
faults but does not sense fault current coming into the generator
from the external system. Placing the CT on the system side of
the generator introduces a problem of the relay not seeing a gen-
erator internal fault when the main breaker is open and when
running the generator isolated from other generation or the util-
ity. If an external source contributes more current than does the
generator, using CTs on the generator terminals, rather than neu-
tral-side CTs, will increase 51 relay sensitivity to internal faults
due to higher current contribution from the external source;
however, the generator is unprotected should a fault occur with
the breaker open or prior to synchronizing.
Voltage-restrained or voltage-controlled timeovercurrent
relays (51VR, 51VC) may be used as shown in Fig. 11 to
remove any concerns about ability to operate before the gener-
ator current drops too low. The voltage feature allows the relays
to be set below rated current. The Basler BE1-951, BE1-1051,
Fig. 9. Phase overcurrent protection (51) must be delayed to coordinate with external BE1-GPS100, and BE1-51/27R voltage restrained approach
relays.
causes the pickup to decrease with decreasing voltage. For
pc_v2 10/19/04 10:25 AM Page 68

Fig. 12. Flux summation relay (50) provides sensitive, high-speed, selective differential
protection (87).
prevent saturation of the CT during internal faults that may tend

to highly overdrive the CT secondary. The 51 relay shown in
Fig. 12 is applied for back-up of external faults and as back-up
for the 50 relay.
The 87G relay in Fig. 13 is connected to respond to phase
differential currents from two sets of CTs. In some applications
it may include a unit differential that includes the step-up trans-
former. In contrast to a 51 or 51V relay that monitors only one
CT, the 87G relay responds to both the generator and external
contributions to a generator fault. Because of the differential
connection, the relay is immune, except for transient CT satura-
tion effects, to operation due to generator load flow or external
faults and, therefore, can provide sensitive, high speed protec-
tion. While the CTs must be of the same ratio, they do not need
to be matched in performance, but the minimum pickup of the
Fig.11. Voltage-restrained or voltage-controlled time-overcurrent phase fault protection. Basler BE1-CDS220 or BE1-87G must be raised as the degree
of performance mismatch increases. (See the BE1- CDS220 and
BE1-87G instruction manuals for specifics on settings.) A min-
example, the relay might be set for about 175% of generator imum pickup of 0.1 times tap (CDS220) or 0.4A (87G) is repre-
rated current with rated voltage applied; at 25% voltage the sentative of a recommended setting for a moderate mismatch in
relay picks up at 25% of the relay setting (1.75*0.25=0.44 times CT quality and burden. Fig. 13 also shows 51V relays to back
rated). The Basler BE1-951, BE1-GPS, and BE1-51/27C volt- up the 87G and external relays and breakers.
age controlled approach inhibits operation until the voltage
drops below a preset voltage. It should be set to function below
about 80% of rated voltage with a current pickup of about 50%
of generator rated. Since the voltage-controlled type has a fixed
pickup, it can be more readily coordinated with external relays
than can the voltage-restrained type. The voltage-controlled
type is recommended since it is easier to coordinate. However,
the voltagerestrained type will be less susceptible to operation
on swings or motor starting conditions that depress the voltage
below the voltagecontrolled undervoltage unit dropout point.
Fig. 12 eliminates concerns about the decay rate of the
generator current by using an instantaneous overcurrent relay
(50) on a flux summation CT, where the CT window can accom-
modate cable from both sides of the generator. The relay does
not respond to generator load current nor to external fault con-
ditions. The instantaneous overcurrent relay (50) acts as a phase
differential relay (87) and provides high-speed sensitive protec-
tion. This approach allows for high sensitivity. For instance, it
would be feasible to sense fault currents as low as 1-5% of gen-
erator full load current. It is common to use 50/5 CTs and to use
1A nominal relaying. A low CT ratio introduces critical satura-
tion concerns (e.g., a 5,000A primary fault may try to drive a
500A secondary on a 50/5 CT). The CT burden must be low to Fig. 13. 87G provides sensitive, high-speed coverage, 51V provides back-up for 87G and
for external relays 87G may wrap step up transformer (unit differential).
pc_v2 10/19/04 10:25 AM Page 69

Fig. 14. Impedance relay, looking for generator and remote line faults.
Another means to detect external faults is with imped-

ance relaying. Impedance relaying divides current by voltage on
a complex number plane (Z = V/I using phasor math) (Figs. 14,
15). Such relaying is inherently faster than time-overcurrent
relaying. In the most common format of impedance relaying,
the tripping zone is the area covered by a "mho" circle on the R-
X plane that has a diameter from the origin (the CT, VT loca- Fig. 17. Reverse-power relay 32-1 prevents load rejection before prime mover shutdown
tion) to some remote set point on the R-X plane. If a fault for selected trips, relay 32-2 operates if motoring is not accompanied by an 86NE operation.
impedance falls within the zone, the relay trips. Multiple zones
may be used, with delays on all zones as appropriate for coordi-
nating with line relays. Impedance relaying is highly direction- In a steam-turbine, the low pressure blades will overheat
al. In Fig. 14, however, because the CT is on the neutral rather with the lack of steam flow. Diesel and gas-turbine units draw
than at the VT, the relay will see faults both in the generator and large amounts of motoring power, with possible mechanical
in the remote system. problems. In the case of diesels, the hazard of a fire and/or
explosion may occur due to unburnt fuel. Therefore, anti-motor-
ing protection is recommended whenever the unit may be con-
nected to a source of motoring power. Where a non-electrical
type of protection is in use, as may be the case with a steam tur-
bine unit, the 32 relay provides a means of supervising this con-
dition to prevent opening the generator breaker before the prime
mover has shut down. Time delay should be set for about 5-30
seconds, providing enough time for the controls to pick up load
upon synchronizingwhen the generator is initially slower than
the system.
Since motoring can occur during a large reactive-power
flow, the real power component needs to be measured at low
power factors. The BE1-32R measures real power down to 0.1
pf. The BE1-951, BE1-1051, and BE1-GPS measure real power
Fig. 15. Impedance relay, looking for remote line faults. down to below 0.01 pf, depending on current magnitude.
Fig. 17 shows the use of two reverse-power relays: 32-1
and 32-2. The 32-1 relay supervises the generator tripping of
REVERSE POWER PROTECTION devices that can wait until the unit begins to motor.
Overspeeding on large steam-turbine units can be prevented by
The reverse-power relay (32) in Fig. 16 senses real power
delaying main and field breaker tripping until motoring occurs
flow into the generator, which will occur if the generator loses
for non-electrical and selected electrical conditions (e.g., loss-
its prime-mover input. Since the generator is not faulted, CTs on
of-field and overtemperature). Relay 32-1 should be delayed
either side of the generator would provide the same measured
maybe 3 seconds, while relay 32-2 should be delayed by maybe
current.
20 seconds. Time delay would be based on generator response
during generator power swings. Relay 32-2 trips directly for
cases of motoring that were not initiated by lockout relay 86NE
— e.g., governor control malfunction.
LOSS-OF-FIELD PROTECTION
Loss of excitation can, to some extent, be sensed within
the excitation system itself by monitoring for loss of field volt-
age or current. For generators that are paralleled to a power sys-
tem, the preferred method is to monitor for loss of field at the
generator terminals. When a generator loses excitation power, it
appears to the system as an inductive load, and the machine
begins to absorb a large amount of VARs. Loss of field may be
detected by monitoring for VAR flow or apparent impedance at
Fig. 16. Anti-motoring (32), loss-of-field (40), protection.
pc_v2 10/19/04 10:25 AM Page 70

THERMAL PROTECTION
Fig. 20 shows the Basler MPS200, BE3-49R, or BE1-49
relay connected to a resistance-temperature detector, embedded
in a stator slot. Relay models are available for either copper or
platinum RTDs. The relay provides a constant-current source to
produce a voltage across the RTD and includes the means to
measure that voltage (proportional to temperature) using sepa-
rate leads. The relays have trip and alarm set points, and the
MPS200 can provide readout of present temperature.
Fig.18. For loss of field the power trajectory moves from point A into the fourth quadrant.
the generator terminals.

The power diagram (P-Q plane) of Fig. 18 shows the
Basler BE1-GPS100 and BE1-40Q characteristic with a repre-
sentative setting, a representative generator thermal capability
curve, and an example of the trajectory following a loss of exci-
tation. The first quadrant of the diagram applies for lagging
power factor operation (generator supplies VARs). The trajecto-
ry starts at point A and moves into the leading power factor zone
(4th quadrant) and can readily exceed the thermal capability of
the unit. A trip delay of about 0.2-0.3 seconds is recommended
to prevent unwanted operation due to other transient conditions. Fig. 20. Stator temperature protection.
A second high speed trip zone might be included for severe
underexcitation conditions.
When impedance relaying is used to sense loss of excita-
tion, the trip zone typically is marked by a mho circle centered LOSS OF VT DETECTION
about the X axis, offset from the R axis by X'd/2. Two zones Two methods in common use to detect loss of VTs are
sometimes are used: a high speed zone and a time delayed zone. voltage balance between two VTs and voltage-current compari-
With complete loss of excitation, the unit will eventually son logic. Fig. 21 shows the use of two sets of VTs on the gen-
operate as an induction generator with a positive slip. Because erator terminals, with the 60FL (Basler BE1-60) comparing the
the unit is running above synchronous speed, excessive currents output of the two VTs. One set supplies the voltage regulator,
can flow in the rotor, resulting in overheating of elements not the other, the relays. If the potential decreases or is lost from VT
designed for such conditions. This heating cannot be detected No. 1, the BE1-60 disables the voltage regulator; if source No.
by thermal relay 49, which is used to detect stator overloads. 2 fails, the BE1-60 blocks relay tripping of the 21, 27, 59N, and
Rotor thermal capability can also be exceeded for a par- 47. In some applications 25, 32, and 40 elements are also
tial reduction in excitation due to an operator error or regulator blocked. Overexcitation relay (24), phase overvoltage (59), and
malfunction. If a unit is initially generating reactive power and
then draws reactive power upon loss of excitation, the reactive
swings can significantly depress the voltage. In addition, the
voltage will oscillate and adversely impact sensitive loads. If the
unit is large compared to the external reactive sources, system
instability can result.
Fig. 19. Loss of excitation using impedance relay Fig. 21. Various voltage protection elements. Voltage-balance relay (60) detects poten-
tial supply failure.
pc_v2 10/19/04 10:25 AM Page 71


Loss of ac potential may also fool the operator into developing
excessive excitation. The 24 relay can only protect for overex-
citation resulting from an erroneous voltage indication if the 24
relay is connected to an ac potential source different than that
used for the regulator.
Fig. 23 shows the relation among the Basler BE1-
GPS100, BE1-951, BE1051, and BE1-24 relay inverse squared
characteristics and an example of a generator and transformer
withstand capability. The generator and transformer manufac-
turers should supply the specific capabilities of these units.
Fig. 22. Loss of fuse, alternate method Phase over (59) and under (27) voltage relaying also acts
as a backup for excitation system problems. Undervoltage relay-
ing also acts as fault detection relaying, because faults tend to
frequency relaying (81), do not need to be blocked, since loss of depress voltage.
potential leads toward non-operation of these functions.
OFF-FREQUENCY OPERATION
OVEREXCITATION AND OVER/UNDER VOLTAGE Diesel engines can be safely operated off normal fre-
PROTECTION quency, and minimal protection is required. Turbine controls
Overexcitation can occur due to higher than rated volt- generally provide protection for off frequency conditions, but
age, or rated or lower voltage at less than rated frequency. For a relaying should be provided to protect the turbine and generator
given flux level, the voltage output of a machine will be propor- during control system failure. Frequency relays are frequently
tional to frequency. Since maximum flux level is designed for applied with steam-turbine units, particularly to minimize tur-
normal frequency and voltage, when a machine is at reduced bine blade fatiguing. IEEE C37.106, Ref. 3 specifically address-
speed, maximum voltage is proportionately reduced. A es abnormal frequency operation and shows typical frequency
volts/hertz relay (24) responds to excitation level as it affects operating limits specified by various generator manufacturers.
thermal stress to the generator (and to any transformer tied to The simplest relay application would be a single underfrequen-
that generator). IEEE C50.13 specifies that a generator should cy stage, but a multiple stage and multiple set point arrangement
continuously withstand 105% of rated excitation at full load. may be advantageous. Each set point may be set to recognize
With the unit off line, and with voltage-regulator control either overfrequency or underfrequency. Multiple frequency set
at reduced frequency, the generator can be overexcited if the points are available in the BE1-81O/U, BE1- GPS100, BE1-
regulator does not include an overexcitation limiter. 951, and BE1-1051.
Overexcitation can also occur, particularly with the unit off line, Another common need for frequency relaying is the
if the regulator is out of service or defective. If voltagebalance detection of generation that has become isolated from the larg-
supervision (60) is not provided and a fuse blows on the regula- er utility system grid. When a generator is connected to the util-
tor ac potential input, the regulator would cause overexcitation. ity, generator frequency is held tightly to system frequency.
Upon islanding, the generator frequency varies considerably as
the governor works to adjust generator power output to local
load. If the generator frequency varies from nominal, islanding
is declared and either the generator is tripped or the point of
common coupling with the utility is opened.
INADVERTENT ENERGIZATION PROTECTION

Inadvertent energization can result from a breaker inter-
rupter flashover or a breaker close initiation while the unit is at
standstill or at low speed. The rapid acceleration can do exten-
sive damage, particularly if the generator is not promptly de-
energized. While relays applied for other purposes may eventu-
ally respond, they are not generally considered dependable for
responding to such an energization.
Figs. 24 and 25 show two methods of detecting the ener-
gization of a machine at standstill or at a speed significantly
lower than rated. This could be caused by single-phase ener-
gization due to breaker-interrupter flashover or 3 phase ener-
gization due to breaker closure. The unit, without excitation,
will accelerate as an induction motor with excessive current
flow in the rotor. Both Fig. 24 and 25 schemes will function
properly with the VT fuses at the generator terminal removed.
With the generator off line, safety requirements may dictate the
removal of these VT fuses. In the case of Fig.24, the overcurrent
protection is enabled by undervoltage units and works as long as
60FL logic does not block the trip path. In Fig. 25 the potential
Fig. 23. Combined generator/transformer overexcitation protection using both the is taken from bus VTs, rather than unit VTs, so the scheme will
inverse squared tripping. Equipment withstand curves are examples only.
pc_v2 10/19/04 10:25 AM Page 72


scheme could be employed where protection independent of the
plant is desired. In this case the 67 relays would be placed in the
switchyard rather than in the control room. While directional
overcurrent relay (67) should be delayed to ride through syn-
chronizing surges, it can still provide fast tripping for generator
faults, since the 67 relays need not be coordinated with external
protection. Fig. 25 shows the operating range for phase A cur-
rent (Ia) with respect to phase B to C voltage (VBC). This range
is fixed by the 60 degree characteristic angle and the ±45 degree
limits set on the operating zone.
NEGATIVE SEQUENCE PROTECTION

Negative sequence stator currents, caused by fault or load
unbalance, induce doublefrequency currents into the rotor that
may eventually overheat elements not designed to be subjected
to such currents. Series unbalances, such as untransposed trans-
mission lines, produce some negative-sequence current (I2)
flow. The most serious series unbalance is an open phase, such
as an open breaker pole. ANSI C50.13-1977 specifies a contin-
uous I2 withstand of 5 to 10% of rated current, depending upon
the size and design of the generator. These values can be
exceeded with an open phase on a heavily-loaded generator. The
Basler BE1-GPS100, BE1-951, BE1-1051, or BE1-46N relay
protects against this condition, providing negative sequence
Fig. 24. Inadvertent enegization protection using instantaneous overcurrent relay (50). inverse-time protection shaped to match the short-time with-
stand capability of the generator should a protracted fault occur.
This is an unlikely event, because other fault sensing relaying
function even if the VT fuses were removed during unit mainte-
tends to clear faults faster, even if primary protection fails.
nance.
Fig. 26 shows the 46 relay connection. CTs on either side
In Fig. 24 the terminal voltage will be zero prior to ener-
of the generator can be used, since the relay protects for events
gization, so the 27 and 81U relay contacts will be closed to ener-
external to the generator. The Basler BE1-46N alarm unit will
gize the timer (62). The instantaneous overcurrent relay (50) trip
alert the operator to the existence of a dangerous condition.
circuit is established after timer 62 operates. Upon inadvertent
Negative sequence voltage (47) protection, while not as
generator energization, the undervoltage and underfrequency
commonly used, is an available means to sense system imbal-
relay contacts may open up due to the sudden application of
ance as well as, in some situations, a misconnection of the gen-
nominal voltage and frequency, but the delayed dropout of 62
erator to a system to which it is being paralleled.
allows relay 50 to initiate tripping. The use of a 60FL function
or two 27 relays on separate VT circuits avoids tripping for a VT
fuse failure. Alternatively, a fuse loss detection or voltagebal-
OUT OF STEP PROTECTION
ance relay (60FL) could be used in conjunction with a single 27 When a generator pulls out of synchronism with the sys-
relay to block tripping. tem, current will rise relatively slowly compared to the instan-
In Fig. 24 the 5 sec pickup delay on timer 62 prevents
tripping for external disturbances that allow dropout of the 27
relays. The 27 relays should be set at 85% voltage (below the
operating level under emergency conditions). The Fig. 25
Fig. 26. Negative-sequence current relay (46) protects against rotor overheating due to a
series unbalance or protected external fault. Negative-sequence voltage relay (47) (less com-
monly applied) also responds.
taneous change in current associated with a fault. The out-of-

step relay uses impedance techniques to sense this condition.
The relay will see an apparent load impedance swing as imped-
ance moves from Zone 1 to Zone 2 (Fig. 27). The time it takes
for the load impedance to traverse from Zone 1 to Zone 2 is used
to decide if an out of step condition is occurring. A moving
impedance is identified as a swing rather than a fault, so appro-
Fig. 25. BE1-67 directional overcurrent relays detect inadvertent energization.
pc_v2 10/19/04 10:25 AM Page 73


Before connecting a generator to the power system, it is
important that the generator and system frequency, voltage mag-
nitude, and phase angle be in alignment, referred to as synchro-
nism checking (25). Typical parameters are shown in Fig. 28.
Typical applications call for no more than 6RPM error, 2% volt-
age magnitude difference, and no more than 10° phase angle
error before closing the breaker. The Basler BE1-951, BE1-
GPS, and BE1-25 all can perform the sync check function.
Auto synchronizing (25A) refers to a system to automat-
ically bring a generator into synchronism with the power sys-
Fig. 27. Out of step relaying (78). tem. It involves sending voltage and speed raise and lower com-
mands to the voltage regulator and prime mover governor.
When the system is in synchronism, the autosync relay is some-
priate fault detection relaying may be blocked. times designed to send a close command in advance of the zero
phase angle error point to compensate for breaker close delays.
SELECTIVE TRIPPING AND SEQUENTIAL TRIPPING The 25 relay, which usually is set to supervise the 25A and man-
It is a practice at some generators to selectively trip the ual sync function, usually is set less tight than the 25A so as to
prime mover, the field, and the generator breaker, depending on coordinate with the actions of the 25A.
the type of fault that is detected. For instance, if the generator is
protected by a 51V and an 87G, and only the 51V trips, it may INTEGRATED APPLICATION EXAMPLES
be assumed that the fault is external to the generator and, hence, Figs. 29 through 33 show examples of protection pack-
the 51V only trips the generator breaker and rapidly pulls back ages.
the excitation governor and prime mover set points. However, if Fig. 29 represents bare-minimum protection, with only
there is no 87G, the 51V trips the entire unit. Associated with overcurrent protection. Generators with such minimum protec-
this concept is sequential tripping used for orderly shutdown. To tion are uncommon in an era of microprocessor-based multi-
prevent overspeeding a generator during shutdown, it is some- function relays. Such protection likely would be seen only on
times the practice first to trip the prime mover and trip the main very small (<50kVA) generators used for standby power that is
breaker and field only after a reverse power relay verifies the never paralleled with the utility grid or other generators. It may
prime mover has stopped providing torque to the generator. appear to be a disadvantage to use CTs on the neutral side as
shown, since the relays may operate faster with CTs on the ter-
SYNCHRONISM CHECK AND AUTO minal side. The increase in speed would be the result of a larg-
SYNCHRONIZING er current contribution from external sources. However, if the
CTs are located on the terminal side of the generator, there will
be no protection prior to putting the machine on line. This is not
recommended, because a generator with an internal fault could
be destroyed when the field is applied.
Fig. 30 shows the suggested minimum protection with
low-resistance grounding. It includes differential protection,
which provides fast, selective response, but differential protec-
tion becomes less common as generator size decreases below
2MVA, on 480V units and below, and on generators that are
never paralleled with other generation. The differential relay
responds to fault contributions from both the generator and the
external system. While the differential relay is fast, the slow
decay of the generator field will cause the generator to continue
feeding current into a fault. However, fast relay operation will
interrupt the externalsource contribution, which may be greater
Fig. 28. Synchronizing parameters: slip, advance and breaker closing time. Fig. 29. Example of bare-minimum protection (low-impedance grounding).
pc_v2 10/19/04 10:25 AM Page 74

Fig. 30. Suggested minimum protection example (low-impedance grounding). Fig. 32. Suggested minimum protection example (high-resistance grounding).
than the generator contribution. Fast disconnection from the unbalance the differential circuit and cause the 87G to trip.
external source allows prompt restoration of normal voltage to Independent CTs could be used to provide improved back-up
loads and may reduce damage and cost of repairs. protection, although this seems to be a minimal advantage here.
The differential relay (87G) may protect for ground However, a separate CT is used for the 51N relay that provides
faults, depending upon the grounding impedance. The 51N protection for the most likely type of fault.
relay in Fig. 30 provides back-up protection for the 87G or will The reverse power relay (32) in Fig. 30 protects the
be the primary protection if the differential relay (87G) is not prime mover against forces from a motored generator and could
sufficiently sensitive to the ground current level. provide important protection for the external system if the
The 51V voltage-controlled or voltage-restrained time motoring power significantly reduces voltage or overloads
overcurrent relay in Fig. 30 is shown on the CT on the high volt- equipment. Likewise, the loss-of-field relay (40) has dual pro-
age/system side of the generator. This allows the relay to see tection benefits—against rotor overheating and against
system contributions to a generator fault. It provides back-up for depressed system voltage due to excessive generator reactive
the differential relay (87G) and for external relays and breakers. absorption. Thermal relay (49) protects against stator overheat-
Since it is monitoring CTs on the system side of the generator, ing due to protracted heavy reactive power demands and loss of
it will not provide any back-up coverage prior to having the unit
on line. If there is no external source, no 87G, or if it is desired
that the 51V provide generator protection while the breaker is
open, connect the 51V to the neutral-side CTs.
Fig. 30 shows three relays sharing the same CTs with a
differential relay. This is practical with solid state and numeric
relays, because their low burden will not significantly degrade
the quality of differential relay protection. The common CT is
not a likely point of failure of all connected relaying. A CT
wiring error or CT short is unlikely to disable both the 87G and
51V relays. Rather, a shorted CT or defective connection will
Fig. 31. Suggested minimum protection example (medium-impedance grounding). Fig. 33. Extended protection example (high-resistance grounding).
pc_v2 10/19/04 10:25 AM Page 75


generator cooling. Even if the excitation system is equipped minals to ground current at the generator neutral. The 51N relay
with a maximum excitation limiter, a failure of the voltage reg- provides backup for the ground differential (87N) and for exter-
ulator or a faulty manual control could cause excessive reactive nal faults, using the current polarizing mode. The polarizing
power output. Frequency relaying (81O/U) protects the genera- winding measures the neutral current.
tor from off nominal frequency operation and senses generator Fig. 32 shows minimum basic protection for a high
islanding. The under and overvoltage function (27/59) detects impedance grounded generator. It differs from Fig. 30 only in
excitation system problems and some protracted fault condi- the ground relay protection and the method of grounding. The
tions. voltage units 59N/27-3N provide the only ground protection,
Fig. 31 shows minimum basic protection for a medium since the ground fault current is too small for phase differential
impedance grounded generator. It differs from Fig. 30 only in relay (87G) operation. The 59N relay will not be selective if
the use of a ground differential relay (87N, part of CDS220 or other generators are in parallel, since all the 59N relays will see
BE1-67N). This protection provides faster clearing of ground a ground fault and nominally operate at the same time. If three
faults where the grounding impedance is too high to sense Phase-Ground Y-Y VTs were applied in Fig. 32, the 27 and 59
ground faults with the phase differential relay (87G). The relay could provide additional ground fault protection, and an addi-
compares ground current seen at the generator high voltage ter- tional generator terminal 59N ground shift relay could be
Fig. 34. BE1-GPS100 applied to low-impedance grounded generator (Low-Z-W25 programmed logic).
pc_v2 10/19/04 10:25 AM Page 76


applied. Fig. 33 shows the application of additional relays for
The Basler BE1-951, BE1-1051, BE1-GPS100, and extended protection: overexcitation relay (24), negative
BE1-59N include a third harmonic undervoltage function (27- sequence overcurrent and overvoltage relay (46 and 47),
3N), that provides supervision of the grounding system, protects ground-overcurrent relay (51GN), voltage-balance relay (60),
for faults near the generator neutral, and detects a shorted or field-ground relay (64F), frequency relay (81) and the 27/50/ 62
open connection in the generator ground connection or in the relay combination for inadvertent energization protection. Relay
distribution transformer secondary circuit. 51GN provides a second means of detecting stator ground faults
Fig. 35. BE1-GPS100 applied to high-impedance grounded generator (HI_Z_GND pre-programmed logic).
pc_v2 10/19/04 10:25 AM Page 77

Fig. 36. BE1-CDS220 applied to generator for 87 phase, 87 neutral, and 51 phase, Fig. 37. Interconnection of BE1-GPS100 and BE1-CDS220, and showing some alterna-
neutral, ground, and negative sequence. tive uses of BE1-GPS100 IG input.
or faults in the generator connections or faults in the delta trans- TYPICAL SETTINGS AND RELAY
former windings. Differential relay 87T and sudden-pressure Table 1 lists the applicable relays discussed herein. The
relay 63 protect the unit step-up transformer. Relay 51N pro- right column provides typical settings for use as a starting point
vides backup for external ground faults and for faults in the in the setting determination procedure. The proper settings are
highvoltage transformer windings and leads. This relay may heavily influenced by the specifics of each application. Typical
also respond to an open phase condition or a breaker-interrupter settings are also used as an aid in selecting the relay range where
flashover that energizes the generator. The 51N relay will be a choice is available.
very slow for the flashover case, since it must be set to coordi- Table 2 lists typical Basler relays applicable to generator
nate with external relays and is a lastresort backup for external protection. There are 3 classes of relays presented in Table 2.
faults. The classical single function "utility grade" (i.e., tested to IEEE
Figure 33 shows wye-connected VTs, appropriate with C37.90 standards) BE1-XXX relays are listed, followed by the
an isolated-phase bus. single function "industrial grade" BE3-XXX relays. (Except the
multifunction BE3-GPR is tested to full IEEE C37.90 stan-
APPLICATION OF NUMERICAL PROGRAMMABLE dards.) Finally, the multifunction utility grade numerical relays
RELAYS are listed.
Numerical programmable relays contain many of the
functions discussed in this guideline in a single package. © 1994, 2001 Basler Electric
Figures 34 through 37 show the BE1-GPS100 and BE1-
CDS220 applied to generator protection. Due to logic complex- Additional information on each relay is available on the
ity, full details are not shown. Details of these applications may Basler Electric web site, www.basler.com.
be found in the respective instruction manual. Updates and additions performed by various Basler
Electric Company employees.
BIBLIOGRAPHY George Rockefeller is a private consultant. He has a BS
1. IEEE C37.101, IEEE Guide for Generator Ground in EE from Lehigh University; MS from New Jersey Institute of
Protection Technology and a MBA from Fairleigh Dickinson University.
2. IEEE C37.102, IEEE Guide for AC Generator
Mr. Rockefeller is a Fellow of IEEE and Past Chairman of IEEE
Protection
3. IEEE C37.106, IEEE Guide for Abnormal Frequency Power Systems Relaying Committee. He holds nine U.S. Patents
Protection for Generating Plants and is co-author of Applied Protective Relaying (1st Edition).
4. J. Lewis Blackburn, “Protective Relaying: Principles Mr. Rockefeller worked for Westinghouse Electric Corporation
and Applications”, 2nd Edition, Marcel Dekker, Inc., 1998. for twenty-one years in application and system design of protec-
5. S. Horowitz and A. Phadke, “Power System tive relaying systems. He worked for Consolidated Edison
Relaying”, John Wiley & Sons, Inc., 1992. Company for ten years as a System Engineer. He has served as
a private consultant since 1982.
pc_v2 10/19/04 10:25 AM Page 78

TABLE 1 - TYPICAL SETTINGS

IEEE No Fig. Function Typical Settings and Remarks
51VR 11 Voltage Restrained Overcurrent PU: 175% INOM; Curve: VI; TD: 4. Zero Restraint Voltage: 100% VNOM L-L
59N: 5% VNEU during HV terminal fault; 27-3N: 25% V3rd during normal opera-
59N, 27-3N, 59P 4 Ground Overvoltage tion; TD: 10s 59P: 80% VNOM
67IE 25 Directional O/C for Inadvertent Energization PU: 75-100% INOM GEN; Definite Time (0.1-0.25 sec.) Inst: 200% INOM GEN
81 21 Over/under frequency Generator protection: 57, 62Hz, 0.5s; Island detection: 59, 61Hz, 0.1s
87G 13 Generator Phase Differential BE1-87G: 0.4A; BE1-CDS220: Min P.U.: 0.1 * Tap; Tap: INOM; Slope: 15%
BE1-CDS220: Min P.U.: 0.1 times tap; Slope 15%; Time delay: 0.1s; choose low tap
87N 8 Generator Ground Differential BE1-67N: current polarization; time: 0.25A; Curve: VI; TD: 2; Instantaneous: dis-
connect
87UD 13 Unit Differential BE1-87T or CDS220 Min PU:0.35*Tap; Tap: INOM; Slope 30%
TABLE 2 - BASLER ELECTRIC RELAY APPLICATION MATRIX
Multifunction (3)
IEEE No Single Function BE1- Single Function BE3-
X=Included •=Optional
BE3-GPR 51 Ph
BE3-GPR 50TN
BE1-1051
BE1-GPS
BE1-CDS
BE1-851
BE1-951
MPS200
24 24 X X X
25 25 25 • • • • •
25A 25A 25A
27 27 27 X X X X X X
27/50IE 50/51B, 50, 27 27. 51 X X X
27/59 27/59 27/59 X X X X X
32 32R 32O/U 32 X X X • • X
40 40Q X • •
46 46N X X X X X X
47 47N 47N X X X • • X
49 49 49R, 49TH X
49/51
50/51G(1) 50/51B, 51 51 X • • • • 50T X
50/51N(2) X X X X
50/87 50/51B, 50 51 X X X X X
51/P 50/51B, 51 51 X X X X X
51VC 51/27C X X X
51VR 51?27R X X X
59P 59 59 X X X
59N 27-3N, 59P 59N •(4) •(4) •(4) X
60FL 60 X X X X X X
67IE 67 X X
81 81O/U 81O/U X X X X X
87G 87G X
87N 87N • •
87UD 87T X
(1) 50/51G - Indicates a relay that monitors a ground CT source.

(2) 50/51N - Indicates a relay that calculates residual (3I0) from phase currents.
(3) Not all functions in relays are shown. Relays also may include multiple set points and setting groups.
(4) BE1-951, -1051, and -GPS have standard capability to calculate 3V0 from wye-connected phase CTs. VAUX input is optional.
pc_v2 10/19/04 10:25 AM Page 79

INTEGRATED PROTECTION, METERING AND

CONTROL USED WITH MONITORING USING WEB
ENABLED COMMUNICATION TECHNOLOGIES
By Ajit Bapat, M.Tech, MBA, P Eng, Schneider Electric Canada
A new dimension is emerging in Electrical Distribution which helps you to avoid costly shut-downs. For example, mon-
systems. The demand is being generated by users who need to itoring the equipment will help identify equipment being
know where electricity is being consumed, identify what is stressed. Quite often power quality problems originate within a
causing power related problems, manage cost of energy, diag- facility. The only way you can determine the cause is with prod-
nose outages or potential outages and to plan for expansion. ucts that can help you monitor, troubleshoot and analyze power
With the installation of modern, particularly non-linear equip- quality problems. By these efforts you will decrease down time
ment, an emphasis is being placed on high efficiency, energy and achieve savings up to 10%.
savings, and a growing environmental concern. These users, of In order to design a distribution system to meet the needs
electrical equipment, need to take proactive steps to maintain of robustness, reliability, cost effectiveness, ease of use and pro-
proper equipment operation. vide the benefits required, the guiding principals are:
We need to monitor electrical systems, record data and 1) Independence of relay protection; To achieve maxi-
minimize downtime by identifying, isolating and correcting mum reliability of protective relay functions in the distribution
problems, enhance facility management, improve operations system, the relay architecture should be such that the protection
and increase value, thereby, allowing us to be alert to potential relay function is held independent of the monitoring, metering
problems before they occur and provide diagnosis and trou- and control function and is provided with highest reliability.
bleshooting capabilities through: device alarm/trip status, log- 2) Open system architecture: The network used for mon-
ging/time and event data logging. In addition, we need to opti- itoring and control must use commonly available hardware and
mize total energy usage and cost allocation to users by taking employ open protocols to keep the cost as low as possible so
into account the tariff structure. Monitoring can assist with the that interoperability with other systems and devices is maxi-
following: optimal loading of electrical circuits, the archiving mized.
and trending of energy information, the appropriate allocation 3) User friendly operator interfaces: The HMI screens or
of energy costs, and the control of generation. Integration of this operator interfaces must be intuitive and easy to use.
equipment in a cost-effective manner allows the informed user 4) Capacity for growth: The system must have capacity
to take proactive, corrective and energy management measures. to grow and expand.
Company environments are constantly changing due to 5) Minimize control wiring hard costs: Hard wiring in the
the pressure of competition and the need for profitability. What field or factory is very expensive. Multidrop communication
is everyone trying to do? Increase productivity and lower costs, should be used to minimize control wiring. Use the most com-
and thus knowing the status of your power system is key to your monly used open architecture of high speed Ethernet. This infra-
achieving these objectives. Getting answers to the questions, structure is used for IT and exists in most facilities. The hard-
“Why did that breaker trip? Who is wasting energy? Where can ware, cables, connectors etc. are commonly available and its
I add this new load?” hasn’t always been easy. Schneider robustness has been well established.
Electric’s Transparent Ready equipment makes using such inte- 6) Integrate with facility management system: Data
gration very cost effective and easy to use and apply. retrieved from the power distribution network is of interest to
In Transparent Ready equipment with integrated network different departments: Operations, Maintenance, Facility,
where are the savings? Finance etc. The information should be presented from the com-
Reduce Energy Costs - Save 2% to 4% on power bills by mon database in a format of use to the user.
knowing where energy is being used. Data is available from What were the Common problems in Integration up to
your entire business to make informed decisions concerning the this point?
entire operation. Such as shifting peak demands, improving Product compatibility - incompatible from different ven-
power factor cost allocating to departments to change behavior dors due to protocols. Many different proprietary protocols
so that waste is avoided. makes the interoperability of communication very difficult.
Increase Equipment Utilization - Save 2% to 5% in oper- Expertise is required to link systems, which makes such systems
ating costs through optimal utilization, identifying spare capac- very expensive for hardware and software.
ity and load balancing thus eliminating unnecessary equipment Maintaining such proprietary systems was expensive and
purchases. This avoids capital costs and maximizes your equip- in addition the first cost was unacceptable. The control panels
ment investment. were not user friendly and not intuitive. Links to the central
Improve reliability – Transparent Ready equipment shared database were too slow. Meaningful easily read and
enables you to find potential problems before they happen interpretive reports were not available. Remote Control access
pc_v2 10/19/04 10:25 AM Page 80


from outside was not secure and systems were not robust. They provide current and voltage measurement with pre-
All these have been overcome by the Transparent Ready cision from 0.04% to 0.1%.
initiative by Schneider Electric. These products are connected to Ethernet via their
It provides easy, simple, affordable access to valuable ECC21 option card. This card also provides an integrated Web
system information using a standard web browser. It is afford- gateway/server function to Modbus on Ethernet TCP/IP for
able, no third-party system integration costs. It has the same Modbus Serial Link (SL) communicating products connected
functionality across an entire family of power and control prod- downstream. The gateway/server function provided by the ECC
ucts from Schneider Electric globally. It uses open industry 21 card is similar to that of the EGX server.
standard communication protocols - Modbus TCP, Modbus. PowerLogic SMS software is a set of software tools spe-
The integrated protection, metering, control and commu- cially designed to help the user of an electrical network to con-
nication system based on universal Ethernet TCP/IP and Web trol this network and reduce the costs connected with electrical
technologies assures interoperability. Every one knows how to power. The tools are deliberately simple to configure, and incor-
launch a web browser in the windows environment. No special porate the library of electrical distribution communicating com-
software is required. Hardware is inexpensive and widely avail- ponents, such as Power Meter, Circuit Monitor, Masterpact or
able. Embedded web pages makes the on-line information avail- Sepam, as standard. They are also open, so that they can handle
able in a user-friendly format very quickly and easily. The any type of Transparent Ready product (using Modbus TCP
industrial automation products and Electrical Distribution prod- messaging).
ucts can be integrated into real-time data-sharing systems, with The software offers a large number of pre-defined
no need for interfaces. screens, enabling users to quickly access the electrical data they
Schneider Electric’s communicating electrical distribu- require. These screens can also be customized. Using such a
tion protection and measurement products are ideal for integra- tool the user can:
tion in Transparent Ready architectures. This range includes in • Ascertain the status of his network.
particular: • View the load curves.
• Sepam series 20, 40 and 80 Medium Voltage (MV) protec- • Optimize his contract with his energy supplier.
tion relays. Provided on Medium voltage switchgear and • Use submetering and thus accurately re-allocate energy
Motor Control. costs.
• Masterpact Low Voltage (LV) circuit breakers used with • Optimize his investment.
their Micrologic A, P or H protection units. Provided in • Monitor electricity quality for optimum operation of the
Low Voltage switchgear and switchboards. network and the process being supplied.
• Power Logic power meters (PM), such as the PM500, The SMS software range covers various needs according
PM700 and PM800 that can be used for both MV and LV. to functional requirements and type of electrical installation.
• The advanced Power Logic Circuit Monitors (CM), such as
the CM2000, CM3000 and CM4000. THE UNIVERSAL COMMUNICATION STANDARD:
Distribution equipment assemblies containing the protec- ETHERNET TCP/IP
tion relays and metering products are connected to Ethernet via The recognition of Ethernet TCP/IP, both in organiza-
a gateway or server that is generally configured for a number of tions and on the Internet, has made it the communication stan-
circuits being monitored in the electrical distribution equipment dard of today. Its wide use is leading to a reduction in connec-
assembly. Typically these gateways have Ethernet connectivity tion costs, increased performance and the addition of new func-
using standard connectors. They can have web addresses and tions, which all combine to ensure its durability. Ethernet
provide web services required for monitoring electrical param- TCP/IP meets the connection requirements of every application:
eters, like measured values, amps, volts, power, energy, wave- • Twisted pair copper cables for simplicity and low cost.
forms, status – open, close, number of ops, temperatures, • Optical fiber for immunity to interference and for long dis-
events, alarms, etc. The monitored parameters are gathered by tances.
these gateways on Modbus serial communication from the indi- • Communication redundancy, inherent in the IP protocol.
vidual devices. The EGX gateway is specifically designed to • Radio or satellite to overcome wiring restrictions.
withstand the harsh thermal and electrical environment in the • Remote point-to-point access via the telephone network or
power distribution equipment. the Internet for the cost of a local call.
CM 3000 and 4000 Circuit Monitors are high-perform- Ethernet TCP/IP, a truly open technology, supports all
ance monitoring units that provide a large number of possibili- types of communication:
ties. Installed on the incoming feeder, on the Medium Voltage • Web pages
(MV) or Low Voltage (LV) switchboard incoming supply or on • File transfer
sensitive outgoing feeder, the Circuit Monitors record the elec- • Industrial messaging, etc
trical parameters of the installation. They provide relevant infor- With its high speed, the network no longer limits the per-
mation for controlling costs, improving power quality and min- formance of the application. The architecture can evolve with-
imizing production downtime. They offer the possibility of out any difficulty. The products remain compatible, ensuring the
detailed recording of the conditions of any detected interfer- long-term durability of the system. Modbus has been the de
ence. facto standard for serial link protocols in industry since 1979. It
Depending on the model and the options, they also per- is used for the communication of millions of automation
form the following: devices. As a result of this success, the Internet community has
• Detection and recording of voltage dips and jumps which reserved the TCP 502 port for Modbus. Modbus can thus be
cause production stoppages on some sensitive processes. used for exchanging automation data on both Ethernet TCP/IP
• Precise (1 ms) time synchronization by GPS. and the Internet, as well as for all other applications (file
pc_v2 10/19/04 10:25 AM Page 81


exchange, Web pages, e-mail, etc). • Provides remote multistation “nomadic” control, without
The simple structure of Modbus is bringing it ever- any special software on the client stations.
increasing success. Users can download the specifications and • Provides a direct link to a company’s information systems,
source code for numerous products that use the Modbus/TCP without the need for an interface.
protocol, free of charge from the Modbus-IDA website
www.modbus-ida.org. TRANSPARENT READY FOR A WORLD WITHOUT
Building on its industrial expertise, Schneider Electric RESTRICTIONS
now has a complete offer of highly user-friendly services on Schneider Electric has a wide range of products:
Ethernet TCP/IP that are dedicated to automation: Modbus TCP Protective Relays, Metering devices, Controllers and PLCs,
messaging, optimized I/O Scanning, publication and subscrip- industrial PCs, HMI devices, variable speed drives, I/O mod-
tion of variables between Controllers and PLCs, automatic ules, gateways, servers, switches, SCADA software, inductive
product reconfiguration, pass band monitoring, system diagnos- identification systems, etc.
tics, etc. These products provide different levels of Web services
The single network, requiring no interfaces between the and communication services on Ethernet TCP/IP, according to
worlds of information technology and automation, is now a real- users’ requirements. In order to simplify choice and ensure their
ity. Schneider Electric broke new ground in 1998 with the first interoperability within a system, each product is now identified
Web servers embedded in automation products. These servers by the class of services it provides.
provide remote access, using a simple Internet browser, to Use - a common Ethernet TCP/IP infrastructure, covering
process information and equipment diagnostics. With all levels, from electrical distribution, control and automation to
FactoryCast HMI, Telemecanique was once again the first in company management.
making the Web servers in Controllers and PLCs “active”. Not Benefit - from the competitive advantages of proven
only does the Web server provide pages containing the system high performance levels.
and process variables, but it also executes tasks totally Reduce - downtime thanks to remote diagnostics via the
autonomously, without making use of the PLC processor: man- Web.
agement of a real-time HMI database, e-mail transmission, cal- Create - secure connections throughout the world.
culations, connectivity with databases, etc. With its functions Minimize - training costs by using tools everyone is
embedded in a PLC, the HMI active Web server: familiar with (Internet browser, etc).
• Simplifies or removes the need for conventional Control - costs by using open universal standards that do
HMI/SCADA (Supervision Control And Data Acquisition) not require any special interfaces.
solutions, reducing communication via polling to update
HMI/SCADA databases.
pc_v2 10/19/04 10:25 AM Page 82


pc_v2 10/19/04 10:25 AM Page 83

A NEW APPROACH TO HIGH IMPEDANCE FAULT

DETECTION
By Ratan Das, Ph.D., P.Eng, ABB
High impedance faults result when an energized primary fire, property damage, and someone coming into contact with
conductor comes into contact with a quasi-insulating object. the live conductor are some of the major concerns.
Typically, a high impedance fault exhibits arcing and flashing at Utilities must always have public safety as a top priority.
the point of contact. The significance of these hard to detect However, high-impedance fault detection has not been possible
faults is that they represent a serious public safety hazard as well in the past and realistic detection algorithms are not anticipated
as a risk of arcing ignition of fires. High impedance fault detec- that can detect 100% of all downed conductors, while having
tion has been a major concern of protective relaying for a long 100% security against misoperation. Utilities need an economic
time. This article discusses some of the techniques (patent pend- solution and a system that can reliably detect high-impedance
ing) that have been implemented in the recently announced fea- faults.
ture of HIF DetectTM in the state-of-the-art distribution feeder Some of the earliest high-impedance fault detection
protective relay from ABB. The article also discusses some of schemes involved arcing fault detection techniques that used
the field trials that have been conducted by performing staged- low frequency current components and are described in [2].
fault testing at different utilities. Reference [3] describes a conventional detection technique that
compares energy values of load currents to preset thresholds in
INTRODUCTION order to detect an arcing fault on a power line. However, this
Electric power systems have grown rapidly over the past patent is very basic in its nature and does not consider or use
fifty years. This has resulted in a large increase in the number other features of load currents to detect a fault.
of lines in operation and their total length. These lines experi- BC Hydro and Powertech Labs Inc. tested three high-
ence faults for many reasons. In most cases, electrical faults impedance fault detection systems [4]. The most significant
manifest in mechanical damage which must be repaired before result was that the higher frequencies of high-impedance fault
returning the line to service. signatures played an important role in high-impedance fault
A vast majority of these faults are ground faults. A con- detection and in distinguishing high-impedance faults from
ventional protection system based on overcurrent, impedance or other types of faults or normal arcing operations. In another
other principles is suitable for detecting relatively low-imped- high-impedance fault detection study [5], the results of years of
ance faults which has relatively large fault current. These con- experience with high-impedance fault detection and testing are
ventional methods have been used with success for a long time summarized and the formal evaluation of the performance of a
to detect such low impedance faults and take necessary action to randomness-based high-impedance fault detection algorithm is
isolate the faulted section of the system. disclosed. This technique was implemented in a product which
However, a small percentage of the ground faults have uses multiple algorithms based mainly on energy, randomness,
very large impedance, often comparable to load impedance, and interharmonics, etc. in layers as described in [6].
consequently have very little fault component of current. There is published literature available on many other
Although these high-impedance faults do not create imminent attempts to detect high-impedance faults. They are reported in
danger to power system equipment, they are a considerable some of the references provided. It can be appreciated that con-
threat to humans and property. The IEEE Power System Relay ventional means for detecting high-impedance faults in electri-
Committee working group on High-Impedance Fault Detection cal power lines are typically not always conclusive and/or reli-
Technology [1] defines High Impedance Faults as those that 'do able and can be expensive. Therefore, a need exists for a new
not produce enough fault current to be detectable by conven- reliable and economic solution for detecting high-impedance
tional overcurrent relays or fuses'. High-impedance fault detec- faults in electrical power lines, which addresses the engineering
tion has been a major concern of protective relaying for a long and legal ramifications of detecting and determining what to do
time. Relaying engineers and researchers have been challenged once a high-impedance fault is detected.
for a long time to develop a suitable technique that will detect This article describes a new approach to detect high-
high-impedance faults with a reasonable degree of reliability, impedance faults [7] and some staged-fault test results [8]. The
while being secure against false detection. approach uses only three algorithms based on higher order sta-
Protective relays are usually designed to protect equip- tistics, wavelets and neural networks, and uses decision logic to
ment (line, transformer, etc.) from damage by isolating the determine the detection of the high-impedance faults.
equipment during high current conditions. High-impedance
faults are typically found on distribution circuits, and resent NEW APPROACH TO HIGH IMPEDANCE FAULT
results in very little current. High-impedance faults do not pose DETECTION
a threat to equipment and by their nature they cannot be detect- The work presented in this article relates to a new
ed with conventional overcurrent devices. Nonetheless, the dan- approach to high-impedance fault detection that includes a
gers of a downed conductor are obvious to all. Possibility of multi-scheme employing individual fault detection systems,
pc_v2 10/19/04 10:25 AM Page 84


order statistical system, a wavelet-based high-impedance fault
detection system, and a neural network-based high-impedance
fault detection system. As shown in Figure 2, power system sig-
nals are acquired, filtered and then processed by individual
high-impedance fault detection algorithm and results of these
individual algorithms are further processed by a decision logic
to provide the detection decision. The decision logic can be
modified depending on application requirement. For example,
one-out-of-three logic will provide maximum dependability,
whereas three-out-of-three logic will provide maximum securi-
ty and two-out-of-three will be somewhere in-between.
Fig. 1. Electrical power distribution network having a composite high-impedance fault
detection system. SIGNAL MODEL
The high-impedance fault detection algorithms reported
each having their own algorithm application that uses various in this article are based on using current signatures in all of the
features of phase and/or ground currents to individually detect a 3-phases and/or ground which are considered non stationary,
high-impedance fault. Suitable features of the currents include temporally varying, and of various burst duration. Even though
their waveform signatures, their sample values etc. Figure 1 the current signals do resemble the third harmonic component
shows a schematic diagram of an electrical power distribution of the current, other harmonic components as well as non-har-
system having an electrical power distribution line and a com- monic components can play a vital role in high-impedance fault
posite high-impedance fault detection system. detection. One challenge is to develop a data model that
The composite high-impedance detection system acknowledges that high-impedance faults could take place at
includes three individual high-impedance fault detection sys- any time within the observation window of the current signal
tems which are not shown in Figure 1 but are shown in Figure and could be delayed randomly and attenuated substantially,
2. Also shown in Figure 1 are the potential transformer PT and depending on the fault location away from the measuring sta-
the current transformer CT which provide the analog inputs for tion. The exemplary model is motivated by previous research,
the high-impedance fault detection system which could be actual experimental observations in the laboratory, and what tra-
implemented in a protective relay used for protecting the line. ditionally represents an accurate depiction of a non stationary
These individual high-impedance fault detection systems signal with time-dependent spectrum.
have individual algorithms for individually detecting high- The high-impedance fault detection problem addressed
impedance faults. These algorithms are based on wavelet, high- in this article is formulated as such:
er order statistics and neural network. The individual high-
impedance fault detection algorithms can each have a different Hypothesis H0 : r(t) = s(t) + n(t) (1)
confidence level. A fault is identified as a high-impedance fault
once it is detected independently by the algorithms and Hypothesis H1 : r(t) = s(t) + n(t) + f(t) (2)
processed through a decision logic.
Figure 2 shows a composite high-impedance fault detec- Where r(t) represents the monitored phase and/or ground
tion system including a higher order statistics-based high- currents. It is assumed that all current recordings are corrupted
impedance fault detection system, identified in Figure 2 as a 2nd with additive Gaussian noise n(t). The high-impedance fault sig-
nature is denoted by f(t) and represents the instantaneous value
of the fault current. Normal load signals are denoted by s(t) and
thus Hypothesis H0 represents a non-fault situation and
Hypothesis H1 represents a high-impedance fault situation.
HIGH-IMPEDANCE FAULT DETECTION USING HIGHER ORDER

STATISTICS-BASED SYSTEM
Figure 3 is a flowchart showing a higher order statistics
based high-impedance fault detection system. Acquired data is
filtered using a bandpass filter. The energy is then calculated
and the calculated energy is compared to a threshold to deter-
mine if a high-impedance fault has occurred.
A detection system and algorithm based on examining
the higher order statistical features of normal currents has been
developed and tested, as discussed below. Higher order spectra,
namely the bispectrum and trispectrum are traditionally recog-
nized as important feature extraction mechanisms that are asso-
ciated with the third - and fourth - order cumulants of random
signals. The bispectrum and the trispectrum are, by definition,
the two-dimensional and three-dimensional Fourier transform
of the third - and fourth - order cumulants defined as,
C 2 (m, n) = E{r (t) r (t + m) r (t + n)} (3)

Fig. 2. Composite high-impedance fault detection system.
pc_v2 10/19/04 10:25 AM Page 85


impedance fault detection, and the scale change keeps track of
bands of frequencies of the current load. Both position and scale
are continuous, therefore, the above transform is not suited for
computation. A discrete version of the transform is needed
which is given by,
⎛k - m⎞
C mn = Σ r(t) ϕ ⎝ ⎯⎯ (5)
k 2n ⎠
where, k, m and n are all integers.
The above transform is implemented by multi resolution

analysis where the signal is decomposed into a low pass and a
high pass component via two separate low pass and high pass
filters known as wavelet decomposition filters. After filtering,
both low pass and high pass signals are down sampled by a fac-
tor of 2. The high pass signal component corresponds to the
first detail look of the signal. The second detail look can be
obtained by further decomposition of the current low pass sig-
nal into two new low pass and high pass components. The third,
fourth, etc. detail signals can be obtained by further decomposi-
tion of subsequent low pass components.
The original signal can be reconstructed with minimal
error from its low pass and high pass components in a reverse
pyramidal manner. It is in these high pass components where
distinct features for high-impedance fault can be located and
distinguished from signatures of other nonlinear loads of tran-
sient and arcing in nature. The decomposition filters are associ-
ated with the type of mother wavelet used.
Fig. 3. Higher order statistic based high-impedance fault detection system. HIGH IMPEDANCE FAULT DETECTION USING NEURAL NETWORK
Figure 5 is a flowchart showing neural network-based
C 3 (m, n, k) = E{r (t) r (t + m) r (t + n) r (t + k)} (4)

where E stands for the expected value.
The algorithm implemented in this study is due in part to

Tugnait [9] and utilizes the integrated polyspectra of single-
phase current loads. The detector is developed such that a detec-
tion decision is made using a second order statistic. This detec-
tor uses additional information beyond energy signatures.
This detector relies on all current spectra including the
in-between harmonics as generated by the pre-processing filter
described earlier.
HIGH-IMPEDANCE FAULT DETECTION USING WAVELET-BASED SYSTEM

Figure 4 is a flowchart showing a wavele-based high-
impedance fault detection system. Acquired data is filtered
using a bandpass filter. Then, as is described in detail below, it
is decomposed in separate high and low pass wavelet decompo-
sition filters. The energy is then calculated and the calculated
energy is compared to a threshold to determine if a high-imped-
ance fault has occurred.
The following is an application of high-impedance fault
detection using a wavelet-based detection system. The continu-
ous wavelet transform of r (t) is
∫-∞r(t) ϕ ⎛t
∞
- p⎞
⎝⎯⎯
C p, s = (5)
s ⎠ dt
where, the wavelet is ϕ(t) , p is the position and s is the
scale.
The position argument keeps track of the temporal

change in current harmonics which is essential to the high- Fig. 4. Wavelet-based high-impedance fault detection system.
pc_v2 10/19/04 10:25 AM Page 86


almost an order of magnitude faster. The best results occurred
when 30 nodes were used in the hidden layer. The network was
trained with 600 cases and had a sum-squared error of 11.8 (8
missed detections and 4 false alarms). Generalization testing on
3600 new inputs resulted in about an 86.06% detection rate with
about a 17.06% false alarm rate. The increased performance of
this network over the previous network is likely due to the
invariance of the frequency spectrum to phase shifts. These per-
formance figures are once again based upon using about 0.5 as
the output threshold for indicating a presence of high-imped-
ance faults. An attempt was made to reduce the false alarm rate
by increasing the output threshold to about 0.75. This resulted
in about an 83.7% detection rate with about a 14.8% false alarm
rate. Increasing the threshold to about 0.95 resulted in about a
77.7% detection rate and about an 11.8% false alarm rate.
Third neural network architecture was a combination of
the two previous networks operating in parallel. If the output of
both networks was greater than 0.5, then a positive detection
decision was indicated. For the cases in which the two neural
networks disagreed as to the presence of a high-impedance fault
current, the output of the two networks was summed and a vari-
able threshold was used to make the decision. A threshold of 1.0
corresponded to making the final decision based upon which
network was more confident in its own decision. For example,
if the output of network 1 was 0.9867 and the output of network
2 was 0.0175, then the sum would be less than 1.0 and no deci-
sion would be made because the output of network 2 is closer to
Fig. 5. Neural network based high-impedance fault detection system. the ideal value of 0 than the output of network 1 is to the ideal
value of 1.
On the other hand, if a more conservative approach were
high-impedance fault detection system. Acquired data is fil- desired in which one chose to reduce the false alarm rate, a larg-
tered using a bandpass filter. The samples are transformed using er threshold approaching 1.5 could be selected. In essence, a
a fast Fourier transform (FFT) which is used only in the second larger threshold gives more weight to the network that indicates
neural network described below, and then mapped into the high- a non high-impedance fault situation.
impedance fault detection plane using the neural network algo- The results indicate that the network using the spectrum
rithm and compared to a threshold to determine if a high-imped- (FFT) of the monitored current is more capable of detecting
ance fault has occurred. high-impedance fault than the network using the actual current
Neural Networks have been successfully used in many samples. Using the sampled current network in tandem with the
applications to solve complex classification problems due to spectrum-based network can reduce the false alarm rates, how-
their ability to create non-linear decision boundaries. The most ever, it doesn't appear to increase the detection rate significant-
common and flexible neural network architecture is the multi- ly. The lack of synchronizing the current's zero-crossing during
layer backpropagation which is constructed from a series of training and generalization may prohibit this neural network
neurons. from detecting some of the patterns or features attributed to
The first neural network investigated used 1000 input high-impedance fault currents, such as asymmetry of half cycles
nodes to the network. No attempt was made to synchronize the and variations from cycle to cycle. The results are encouraging
zero crossing of the monitored current to the first input node of given that the detection is performed on only a 3-cycle snapshot
the neural network with the hopes of reducing implementation of data.
complexity. The best results occurred when using 200 nodes in
the hidden layer. The network was trained with 600 input/target LABORATORY TEST RESULTS
cases (300 high-impedance fault data and 300 load data) and Figure 6 shows in simplified form an exemplary labora-
had a sum-squared error of 1.4 after completion of learning (1 tory model that was developed to experimentally stage high-
missed detection and 0 false alarms). Generalization perform- impedance faults and to collect data for testing and evaluation.
ance was determined by testing the network on 3600 new 3 The exemplary setup included two 120/4500 V, 1 kVA trans-
cycle windows (1800 high-impedance faults and 1800 loads). formers connected in parallel and energized from a 120 V, 15 A,
Considering all network output greater than 0.5 to indicate the 60 Hz power source. As shown in Figure 3, a bare conductor
presence of a high-impedance fault, the network achieved a was connected to one terminal of the transformer secondaries to
detection rate of 70.83% with a 22.06% false alarm rate. simulate a downed transmission line. The other secondary ter-
The second neural network design used the spectrum of minal was connected to a copper plate buried in soil, thereby
the 3-cycle window of data. The magnitude of the FFT of the simulating the ground electrode and the earth.
1000 samples was truncated at the 13th harmonic. This resulted The bare conductor was dropped on a variety of soil sur-
in a reduction to only 40 input nodes for the neural network. faces to investigate differences in the resulting currents. The
This network had fewer weights and biases and could be trained current signatures were collected using a data acquisition sys-
pc_v2 10/19/04 10:25 AM Page 87

Fig. 6. Laboratory model developed to experimentally stage high-impedance faults.
tem based on the National Instruments data acquisition and sig-

nal conditioning boards with Lab-VIEW software operating on
Windows NT. The data was sampled at 20 kHz, quantized to 14 Fig. 7. HIF detection units along with the data collection system.
bits and stored in binary format. Each trial case was conducted
for a 50-second duration.
Fifteen cases were run for seven different wet surface separate data acquisition system for the HIF staged-fault testing
conditions (wet and frozen sod, soil, asphalt, gravel, sand, and was also developed as shown in Figure 7. This has been done
concrete) for a total of 105 high-impedance fault cases. This to collect data from the staged-fault testing, independent of the
data acquisition scheme was also used to collect signatures for HIF detection units, for future use and replay of the fault events.
currents for single-phase nonlinear loads (e.g., TV, fluorescent The data acquisition system is based on National Instruments
lamp, PC, bridge rectifier, a phase-controlled motor drive, and Labview PC software using off-the-shelf Data Acquisition
arc welder). A total of 22 load files were created. System (DAS) card and other hardware from National
The higher order statistics-based algorithm was tested Instruments. Assembly of the data collection system and devel-
using the collected data and the results indicate a probability of opment of the software for the system was done at ABB. The
detection of about 97.14% with a zero false alarm rate for all sampling frequency used for data acquisition is about 20 kHz.
cases which includes two arc welder loads. Thresholds were set
such that the false alarm rates αs = 0.05 and αh = 0.05 which FIELD TRIALS
corresponds to an overall false alarm rate of about 0.09. These Since most Protective Relay Engineers spend their time
results indicate that higher order signatures are distinguishable in minimizing the effect of faults on the electric system, they are
from welding and other non-linear loads. not usually comfortable with the process of staging faults on a
The wavelet-based algorithm delivers about 80% detec- system in service. However, utilities have developed a method
tion with about a 0.5% false alarm rate in the absence of arc
welding loads. With the lowering of thresholds, the detection
rate increases to about 95% with about a 0.1% false alarm rate.
The detection performance drops to about 65% in the presence
of arc welding signals and without lowering the thresholds. The
false alarm rate remains under about 1%.
PROTOTYPE HIF DETECTION UNIT AND DATA

COLLECTION SYSTEM
High impedance fault detection algorithms discussed
above have been extensively tested between 1998 and 2000
using data generated at the laboratory. The laboratory results
encouraged ABB to explore the possibility of implementing the
techniques in an embedded platform so that the benefits of HIF
detection can be passed on to the users of protective relays. A
project was undertaken in 2000 to this effect which also includ-
ed the testing of the implemented algorithms with additional
field data. Additional HIF field data was obtained in 2002 from
a research laboratory that independently performed staged-fault
testing in a distribution system. The implemented system
required adaptation and modification to perform satisfactorily
both with laboratory data and acquired field data.
Once satisfied with the performance of the implemented
HIF algorithms, with laboratory and acquired field data, ABB
approached some utilities to verify the performance of the unit
and collection of HIF data with staged-fault testing. Many util-
ities responded to the request positively and successful staged-
fault testing has been conducted.
In addition to the two prototype HIF detection units, a
Fig. 8. Validation of HIF Detect™ on gravel and sand
pc_v2 10/19/04 10:25 AM Page 88


Engineering from the University of
Saskatchewan, Canada. From 1981 to
1992 he worked for NTPC in the area of
power systems relaying and control. He
holds two patents, and authored and co-
authored papers including an invited arti-
cle for John Wiley's online Encyclopedia of
Electrical and Electronics Engineering. He
is a senior member of IEEE, a member of
IEEE Power Systems Relaying Committee,
a member of CIGRÉ and a registered pro-
fessional engineer in Saskatchewan,
Canada. His areas of interest include sim-
ulation of power systems, and protective
relaying algorithms, applications and test-
ing with special interest in fault location
and fault detection techniques.
REFERENCES
[1] IEEE Power System Relaying
Committee Working Group D15 Report,
“High-Impedance Fault Detection
Fig. 9. Validation of HIF Detect™ on concrete and grass. Fig. 10. Validation of HIF Detect™ on asphalt and trees. Technology”, http://www.pes-psrc.org/,
(Go to Published Reports, then to Reports
of staging faults that is both safe for utility personnel perform- Relating to Line Protection - File name is D15MSW60.htm).
ing the tests and results in no interruption or break in service to [2] B. Russell, K. Mehta, R. Chinchali, “An arcing fault
the utility consumers connected to the distribution feeder. detection technique using low frequency current components
Field trial of the HIF Unit was done in the process of col- performance evaluation using recorded data”, IEEE
lecting data from staged HIF testing – these trials took place Transactions on Power Delivery, 1988, Vol. 3, No. 4, pp. 1493-
four times between January and September 2004 at various 1500, October.
locations. At the time of writing, one more field trial is planned [3] B. Don Russell and B. Michael Aucoin, “Energy
at the end of October 2004. Analysis Fault Detection System”, 1996, U.S. Patent No.
Photographs of staged-fault testing on various surfaces 5,512,832.
are shown in Figures 8 through Figure 10. [4] V. Bucholz, M. Nagpal, J. Nielson, B. Parsi, and W.
Results of filed trials were very encouraging and ABB Zarecki, “High impedance fault detection device tester”, IEEE
decided to implement the HIF Detect™ as a standard feature in transactions on Power Delivery, 1996, Vol. 11, No. 1, pp. 184-
their state-of-the-art protective relay REF 550. The user can 190, January.
select the level of security in the HIF notification with a very [5] C. Benner and B. Russell, “Practical high-impedance
intuitive setting called HIF level – set anywhere between 1 and fault detection on distribution feeders”, IEEE Transactions on
10 in steps of 1, 10 being more secure than 1. This is the only Power Delivery, 1997, Vol. 33, No. 3, pp. 635-640, May/June.
setting available for the HIF Detect™ feature, with the factory [6] R. Patterson, W. Tyska, B. Don Russell and B.
default setting being 5. Michael Aucoin, “A Microprocessor-Based Digital Feeder
Monitor With High-Impedance Fault Detection”, presented to
CONCLUSIONS the Forty-Seventh Annual Conference for Protective Relay
- Requirements of High Impedance Fault (HIF) Detection is Engineers, Texas A&M University, 1994, College Station, TX,
different from that of conventional protection or relaying. March 21-23.
- Reliable detection of HIF provides safety to human and ani- [7] R. Das and S. A. Kunsman, “A Novel Approach for
mal life, prevent fire and minimize property damage. Ground Fault Detection”, presented to the Fifty-Seventh Annual
- Innovative technology for High Impedance Detection is Conference for Protective Relay Engineers, Texas A&M
available and discussed. University, College Station, TX, 2004, March 30-April 1.
- Results of many successful field trials indicate that the [8] M. Carpenter, R. Hoad, T. Bruton, R. Das, S. A.
developed technology works as expected. Kunsman and J. Peterson, “Staged-Fault Testing for High
- The developed technology is safe, secure, dependable and Impedance Fault Data Collection”, presented to the Thirty-First
freely available with the IED or Relay for feeder protection. Annual Protective Relay Conference, Spokane, WA, 2004,
- The notification from the HIF Detect™ can be used for tak- October 19-21.
ing appropriate action as desired by the user. [9] J. Tugnait, "Detection of Random Signals by
Integrated Polyspectral Analysis", IEEE Transactions on Signal
Ratan Das is the Principal Engineer at ABB in Allentown, PA. processing, 1996, Vol. 44, No. 8, pp. 2102-2108,
He joined the company in 1998. He received his BEE degree
with honors from Jadavpur University, India in 1982 and his
M.Sc.(1995) and Ph.D. (1998) degrees in Electrical
pc_v2 10/19/04 10:25 AM Page 89

A LOOK AT FUSEOLOGY
Written by Tim Crnko, Cooper Bussman
NEED FOR OVERCURRENT PROTECTION the system. A choice is made as to the type of protection,
The purpose of overcurrent protective devices, such as whether fuses or circuit breakers, when a system is
fuses and circuit breakers, is to protect electrical circuit compo- designed/installed or partially upgraded. There are considerable
nents and equipment from the effects of overcurrents. An over- performance and life cycle differences between fuses and circuit
current can be either an overload current or a fault current. An breakers. The building owner must live with the selection of the
overload current is an excessive current relative to normal oper- overcurrent protective device for the life of the electrical system
ating current, and the current stays in the normal conductive or until major (costly) alterations are made. Whether it is the
path. For example, when a motor is mechanically overloaded first day of the electrical system or many years later, it is impor-
beyond its HP rating, it draws current above its rating. Overload tant that overcurrent protective devices perform as intended
currents increase the temperature of conductors and other com- under overload or fault conditions. Modern current-limiting
ponents. If overloads are not interrupted in a prescribed time, fuses operate by very simple, reliable principles and do not
the thermal effects can result in insulation deterioration. require periodic maintenance. Circuit breakers, on the other
Improper overload protection may eventually result in ignition hand, are mechanical devices and there-
of the insulation or adjacent materials or may result in a short- fore, if used, must be periodically main-
circuit. tained and/or tested as per the manufac-
In contrast, a fault causes current to flow outside the nor- turer’s recommendations.
mal conducting path such as might occur when a wrench is
dropped across the bare buses of an energized switchboard. A VOLTAGE RATING
short-circuit or fault current can be many hundred times larger The voltage rating is extremely
than the normal operating current. A high level fault may be important for overcurrent protective
50,000 amperes or larger. If not cut off within a matter of a few devices. The proper application of an
thousandths of a second, extensive damage and violent destruc- overcurrent protective device requires
tion can result – short-circuit currents can cause severe insula- that the voltage rating of the device be
tion damage, vaporization of conductors, arcing, fires, and rup- equal to or greater than the system volt-
turing of equipment. age. When an overcurrent protective
For safety of property and people, overcurrent protection device is applied beyond its voltage rat- Fig. 1. This photograph
ing, there may not be any initial indica- vividly illustrates the effects
is one of the top two most important electrical system aspects;
tors. Adverse consequences typically of overcurrents on electrical
the other is proper grounding. In any given system the frequen-
result when an improperly voltage rated components when protective
cy of overloads and faults can vary widely. Electrical distribu-
device attempts to interrupt an overcur- devices are not sized to the
tion systems are often quite complicated, and they cannot be
rent, at which point it may self-destruct ampere rating of the compo-
absolutely fail-safe. Harsh environments, general deterioration,
in an unsafe manner. nent.
accidental damage or damage from natural causes, human error,
excessive expansion, lack of maintenance, improper mainte- There are two types of AC volt-
nance or overloading of the electrical distribution system are age ratings for overcurrent protective
factors which contribute to the occurrence of such overcurrents. device: straight voltage rated and slash
Whether the occurrences of overcurrents are rare or frequent, voltage rated. The proper application is
the fact is overcurrents do happen and overcurrent protection is straightforward for overcurrent protec-
a necessity. Also, some people debate that bolted fault currents tive devices with a straight voltage rat-
seldom occur; implying why protect against the worst-case fault ing (i.e. 600V, 480V, 240V, etc.) which
situation. Overcurrent protection is similar to insurance. You have been evaluated for proper perform-
hope you never need it, but if and when you do need it, you are ance with full phase-to-phase voltage
glad you have full coverage. Knowing overcurrent protection is used during the testing, listing and
a critical aspect to electrical systems, select the most reliable marking. For instance, all fuses are
overcurrent protective devices and consider the many critical straight voltage rated and there is no
aspects mentioned in this article. need to be concerned about slash rat-
ings. However, some mechanical over- Fig. 2. Considerable dam-
RELIABILITY current protective devices are slash voltage to electrical equipment
age rated (i.e. 480/277V, 240/120V, can result if the interrupting
Overcurrent protection must be reliable. Selecting the
600/347V, etc.). Slash voltage rated rating of a protective device
type of overcurrent protective device to be used should take into
devices are limited in their applications is inadequate and is exceeded
account reliability of the overcurrent protection for the life of
and extra evaluation is required when by a short-circuit current.
pc_v2 10/19/04 10:25 AM Page 90


they are being considered for use.
Most low voltage power distribution fuses have 250V or
600V ratings (other ratings are 125V, 300V, and 480V). The
voltage rating of a fuse must be at least equal to or greater than
the circuit voltage. It can be higher but never lower. For
instance, a 600V fuse can be used in a 208V circuit. The voltage
rating of a fuse is a function of its capability to open a circuit
under an overcurrent.
AMPERE RATING Without Selective Coordination With Selective Coordination

Every fuse has a specific ampere rating. In selecting the
ampere rating of a fuse, consideration must be given to the type
of load and code requirements. The ampere rating of a fuse nor- Fig. 3. Selective Coordination: Avoids Blackouts
mally should not exceed the current carrying capacity of the cir-
cuit. For instance, if a conductor is rated 20 amperes, a 20 tive coordination has been achieved when, in fact, it has not.
ampere fuse is the largest that should be used. As a rule, the Selective coordination can be defined as the act of isolat-
ampere rating of a fuse and switch combination should be ing a faulted circuit from the remainder of the electrical system,
selected at 125% of the continuous load current (this usually thereby eliminating unnecessary power outages. The faulted cir-
corresponds to the circuit capacity, which is also selected at cuit is isolated by the selective operation of only that overcur-
125% of the load current). There are exceptions, such as when rent protective device closest to the overcurrent condition.
the fuse-switch combination is approved for continuous opera- Modern fuses are easy to apply so that selective coordi-
tion at 100% of its rating. nation is designed-in the system. By maintaining a minimum
However, there are some specific circumstances in which ratio of fuse ampere ratings between an upstream and down-
the ampere rating is permitted to be greater than the current car- stream fuse, selective coordination is achieved. Each fuse man-
rying capacity of the circuit. For example in motor circuits, ufacturer publishes minimum selectivity ratios for each fuse.
dual-element fuses generally are permitted to be sized up to Below is an application example where the published minimum
175% and non-time-delay fuses up to 300% of the motor full- fuse selectivity ratio is 2:1
load amperes. If the motor cannot start at this sizing then the
size may even be increased based on some limits in the National
Electrical Code.
INTERRUPTING RATING
Most people in the industry have some knowledge of
voltage rating and ampere rating, but many do not understand
nor take the steps to properly apply overcurrent protective Fig. 4. This diagram shows the minimum ratios of the ampere ratings of LOW-PEAK* YEL-
devices in regards to interrupting rating. This can be a serious LOW fuses that are required to provide "selective coordination" (discrimination) of upstream
property and human safety issue. Interrupting rating is the max- and downstream fuses.
imum short-circuit current that a fuse or circuit breaker has been
tested to interrupt under specified test conditions. It is critical
that a fuse or circuit breaker be able to withstand the destructive
energy of any short-circuit current it may be called upon to CURRENT-LIMITATION
interrupt. If a fault current exceeds the interrupting rating of an If a protective device cuts off a short-circuit current in
overcurrent protective device, the device may actually rupture less than one-half cycle, before the current reaches its total
causing additional damage. Therefore, it is important when available (and highly destructive) value, the device limits the
applying a fuse or circuit breaker to use one which can sustain current. Many modern fuses are current-limiting. They restrict
the largest potential short-circuit current available at the point of fault currents to such low values that a high degree of protection
installation. The interrupting rating of most branch-circuit, is given to circuit components against even very high short-cir-
molded case, circuit breakers typically used in residential serv- cuit currents. If not limited, short-circuit currents can reach lev-
ice entrance panels is 10,000 amperes. Larger, more expensive els of 30,000 or 40,000 amperes or higher (even above 200,000
circuit breakers may have interrupting ratings of 14,000 amperes) in the first half cycle (0.008 seconds at 60 Hz) after
amperes or higher. In contrast, most modern, current-limiting the start of a short circuit. The heat that can be produced in cir-
fuses have an interrupting rating of 200,000 or 300,000 amperes cuit components by the immense energy of short-circuit cur-
and are commonly used to protect the lower rated circuit break- rents can cause severe insulation damage or even explosion. At
ers. The Code requires equipment intended to break current at the same time, huge magnetic forces developed between con-
fault levels to have an interrupting rating sufficient for the cur- ductors can crack insulators and distort or destroy bracing struc-
rent that must be interrupted. tures. Thus, it is important that a protective device limit fault
currents before they reach their full potential. Modern fuses are
SELECTIVE COORDINATION current-limiting due to their inherent design.
Selective coordination can be a critical aspect for electri-
cal systems. Quite often in the design phase or equipment selec- COMPONENT PROTECTION
tion phase, it is ignored or overlooked. And when it is evaluat- Choosing overcurrent protective devices strictly on the
ed many people misinterpret the information thinking that selec- basis of voltage, current, and interrupting rating alone will not
pc_v2 10/19/04 10:25 AM Page 91

Fig. 5. A non-current protective device, by permitting a short-circuit current to build up to Fig. 6. In its current-limiting range, a current-limiting fuse has such a high speed of
its full value, can let an immense amount of destructive short-circuit heat energy through response that it cuts off a short-circuit long before it can build up to its full peak value.
before opening the circuit.
assure protection of circuit component from short-circuit cur- able short-circuit current and the overcurrent protective device
rents. Much of the code is merely a cookbook for matching the let-through characteristics. When the available short-circuit cur-
ampere rating of conductors, equipment and overcurrent protec- rent exceeds a components withstand, it is imperative that the
tive devices. Merely matching the ampere rating of a component current-limiting overcurrent protective let-through current be
with the ampere rating of a protective device will not assure less than the component withstand. Proper protection of circuits
component protection under short-circuit conditions. Also, the will improve reliability and reduce the possibility of injury.
interrupting rating of a protective device pertains only to that Electrical systems can be destroyed if the overcurrent devices
device and has absolutely no bearing on its ability to protect do not limit the short-circuit current to within the withstand rat-
connected downstream components. Quite often, an improperly ing of the system’s components.
protected component is completely destroyed under short circuit
conditions while the protective device is opening the faulted cir- This article gave a brief overview of some critical aspects for
cuit. proper overcurrent protection selection. For more information
There is not sufficient space in this article to show the visit www.bussmann.com. The Bussmann® SPD Selecting
steps to achieve proper component protection, so just an Protection Devices handbook can be viewed or downloaded in
overview is given. Proper component protection requires analy- whole or in sections. Plus there is an assortment of application
sis of the circuit components’ short-circuit withstand, the avail- materials and tools at his site.
pc_v2 10/19/04 10:25 AM Page 92


pc_v2 10/19/04 10:25 AM Page 93

BUYERS GUIDE
ABLOY CANADA INC CD Nova & Associates Co.

9500 Trans Canada 5330 Imperial Street
Montreal, QC H4S 1R7 Burnaby, BC V5J 1E6
Tel: 514-335-9500 Tel: 604-430-5612
Fax: 514-335-0430 Tel: 800-663-0615
Company Description: ABLOY manufactures a complete Fax: 604-437-1036
range of high security locks to fit virtually all applications. Company Description: Exclusive distributor for
ABLOY Padlocks and re-useable Demand Meter Seals are in power/energy products serving electric utility applications.
use extensively in the Electric Utility Industry. These locks offer Teleprotection, Protective relaying systems, Data
the maximum in high security as well as excellent resistance to Communication and Spread Spectrum Radio systems,
harsh environmental conditions. Features include the patented Telemetry, Fault and Event recorders, Power supply systems.
pick Resistant patented keys with signature protection along
with exclusive reserved codes. Also Hardened Boron shackles
with toe and heel deadlocking, no spring loaded parts to years, CITEL, INC.
immediate service. ISO 9001 Quality rating A reference list of 1515 NW 167th Street, Suite5-223
Electric Utilities using ABLOY Padlocks available upon Miami, Florida 33169
request. Tel: (305) 621-0022
Please call toll Free number 1-800-465-5761 for free Fax: (305) 621-0766
ABLOY brochures. Email: citel4u@ix.netcom.com
www.citelprotection.com
Description of products/services: Citel manufactures surge
suppressors for AC Power, Telephone and Data Lines.
They are ideal to protect PBXs, MUXs, PLCs, RTUs and
SCADA systems against lightning surges and electrical tran-
sients.
Many models are offered for RS232, RS422,RS423,
AMETEK SOLIDSTATE CONTROLS RS485, T1/E1, ISDN, 56k/DDS, leased lines, telephone lines;
875 Dearborn Drive for coaxial cable and twisted pairs.
Columbus, Ohio 43085
Tel: (614)846-7500
Fax:(614)885-3990 DUNCAN INSTRUMENTS CANADA LTD.
Email:nick.yarnell@ameteksci.com 121 Milvan Drive
Description of products/services: Design and manufacture Toronto, Ontario M9L 1Z8
custom Industrial-Grade PWM & Ferroresonant and Tel: 416 742-4448
Uninterruptible Power Supply systems, and related electrical Fax:416 749-5053
distribution equipment (Inverters, Battery Chargers, Voltage Email: sales@duncaninstr.com
Regulators, Power Conditioners, Distribution Panels and www.duncaninstr.com
Switchboards). Description of products/services: Duncan Instruments
Also manufacture Automatic Switching Rectifiers for Canada is a leading manufacturers' representative and master
Cathodic Protection systems. distributor for a wide range of utility and electrical instrumenta-
tion.
We can offer you data loggers, power line analyzers,
pc_v2 10/19/04 10:25 AM Page 94


power/energy/harmonics analyzers, power disturbance monitors
and fused test leads/accessories.
YOU NEED In addition to sales, Duncan Instruments Canada can also

provide: calibration - traceable to NRC, technical product sup-
port and application training, instrument repair/modifications,
and rental of selected electrical instruments. Registered to ISO
POWER, 9001:2000.
we provide
power quality FERRAZ SHAWMUT CANADA INC.
corrections. 88 Horner Avenue
Toronto, Ontairo M8Z 5Y3
Ph: 416 -252-9371
Gentec’s optimized global solutions in Fax : 416 252-6572
energy management result in superior Ferraz Shawmut – the world’s premier producer of industri-
quality equipment supported by world- al fuses and the one truly global provider of circuit protection
for industry, with products ranging from low-voltage to medi-
renovated expertise and proven by the um-voltage fuses and accessories, low- and high-power switch-
most efficient network in North America. es and value engineered products to meet most customer
Undeniably, Gentec can achieve the best demands. With over 125 years of experience as a provider in the
return on your investment. For 40 years, circuit protection industry, Ferraz Shawmut is truly your circuit
protection resource.
Gentec has been your best partner in
energy management and the most
cost-effective for your network. FLEXCORE
6625 McVey Blvd.
Contact us and discover our complete Columbus, Ohio 43026
Tel: 614-889-6152
range of equipment and solutions for Fax: 614-876-8538
reactive power: Company Description: FLEX-CORE provides split-core,
• Low and High Voltage Units solid-core, clamp on current transformers, potential transform-
• Automatic P.F. Correction ers, electrical transducers for AC and DC applications, harmon-
ic analyzers, signal conditioners, panel meters, KWH sub-
• Low Voltage Harmonic Filters
metering systems and analog to pulse converters.
• Medium Voltage P.F. Correction
• High voltage harmonic Filters
• Static Var Compensators Systems
• Electronic P.F. Controllers
GENTEC INC.
GLOBAL SOLUTIONS IN
2625 Dalton
ENERGY MANAGEMENT Ste-Foy, Quebec G1P-3S9
Tel:(418) 651-8001
WORLD HEADQUARTERS Fax: (418) 651-2937
2625 Dalton, Sainte-Foy, Québec, Canada G1P 3S9 www.gentec.ca
Tel: 1-800-463-4480 Ext. 237 • Fax: (418) 651-6695 Description of products/services: Gentec offers the most
complete range of power quality solutions in:
US HEADQUARTERS
- Power Factor Correction in Automatic and Fast
35 Gateway Drive, Suite 201, Plattsburg, NY 12901
Compensation system
Tel: 1-888-235-7506 • Fax: (518) 793-2687 - Low- and High-voltage Harmonic filter, Low- and High-
ONTARIO/WESTERN CANADA voltage Reactive filter
614 Elm St., St-Thomas, Ontario, Canada N5R 1K7 - Low-voltage Active filter
Tel: (519) 637-0817 • Fax: (519) 637-1237 - Static Var Compensation System
- Load Battery Charger, Inverter and UPS System.
pc_v2 10/19/04 10:25 AM Page 95

G.T.WOOD CO.LTD.
3354 Mavis Road MTE CORPORATION
Mississauga, Ontario L5C 1T8 W147 N9525 Held Drive
Tel: (905) 272-1096 Menomonee Falls, Wisconsin 53051
Fax: (905) 272-1425 Tel: (262) 253-8200
Email: Isnow@gtwood.com Fax (262) 253-8222
www.gtwood.com Email: sales@mtecorp.com
Description of products/services: Specializing in high volt- Website: www.mtecorp.com
age electrical testing, inspection, maintenance and repairs. Description of products/services: Electrical Power Quality
Supply, refurbishing and repair of new and reconditioned trans- Products including: Harmonic Filters (Guaranteed
formers, structures, switchgear and associated equipment. Performance), AC Line/Load Reactors, RFI Filters, Motor
Infrared thermography, engineering studies and PCB manage- Protection Filters and PQ Transformers.
ment
SCHWEITZER ENGINEERING LABORATORIES, INC.

2350 N.E. Hopkins Court
Pullman, WA 99163
Tel: 519-3321890
Fax: 519-332-7990
Company Description: SEL serves electric power utilities
and industrial companies worldwide through the design, manu-
facture, supply, support of products and services for electrical
HUBBELL CANADA INC. - POWER SYSTEMS system protection, control, monitoring, integration, and automa-
870 Brock Road South tion.
Pickering, Ontario L1W 1Z8 SEL exists to make electric power safer, more reliable, and
Tel: (905) 839-4332 more economical.
Fax: (905) 893-2069
www.hubbellonline.com WINTEK ENGINEERING LOG0
Description of products/services: Pick Up from
WIDE RANGE OF ELECTRICAL EQUIPMENT for the Power Quality and Grounding Handbook -
power industry including: ground set testers; surge arresters; volume 5 page 117
grounding equipment; VERSACRIMP® dieless compression
tools; cable accessories including the Hubbell SAFE-T-RING
device used to eliminate flashovers caused by partial vacuum. WINTEK ENGINEERING LTD.
90 Rankin Street
Waterloo, Ontario N2V 2B3
Tel : (519) 884-7999 or 1866-946-8351
Fax : (519) 884-5333
Email wintek@wintek-eng.com
Website:wintek-eng.com
INTERNATIONAL Description of products/services: Power Distribution

INNOVATIVE SYSTEMS Design and Analysis.
934 Queen Street East, Suite 1 - Industrial
Sault Ste Marie, Ontario P6A 2B9 - Commercial
- Institutional
Tel:(705) 759-4862
Load Studies, One-Line development, Power Quality mon-
Fax:(705) 759-7083 itoring, analysis, troubleshooting, and correction.
Email: admin@no-surge.com
www.no-surge.com
Description of products/services: Sales and distribution of
power quality related products.
4colourPages 10/19/04 10:37 AM Page 2

Make the right

energy decisions
at the right time
with graphical energy dashboards
You'll make faster, more accurate decisions about your energy with ION® meters and software
from Power Measurement. With better insight into your cost equations, you'll slash energy costs
and boost uptime. Use graphical energy dashboards to see exactly what's happening with your
operations and performance. Data that took weeks to analyze now takes seconds. And you can
© 2004 Power Measurement.
access it all immediately, wherever you are, through email or web pages, from your laptop,
cell phone, pager, or PDA.
Beyond off-the-shelf ION systems, you can also access custom solutions through our
comprehensive range of engineering services, including specialized user interfaces, network
architecture designs, and third-party integration. Save months of study and design effort by
capitalizing on our expertise to quickly turn your business goals into high-performance solutions.
ION is a registered trademark of Power Measurement.
Call us today for a free metering and software System Design Handbook.
www.pwrm.com Toll free 1-866-ION-SMART (866-466-7627)
smart energy everywhere

4colourPages 10/19/04 10:37 AM Page 4

E-Factor
E N E R GY S AV I N G S 104-974 Queen St. E.
Sault Ste. Marie
ON P6A 2C5
E-Factor Transformer Product Benefits Tel: (705) 759-4862
Fax: (705) 759-7083
Saving energy means saving costs. E-Factor equipment is number one at both. E-Mail: admin@no-surge.com
http://www.e-factor.ca/
1. The initial payback period after a first-cost purchase is the fastest in the industry.
2. With proven efficiency and long-term durability that is certified and third-party validated...
the savings compound for a continuous, increasing return-on-investment.
3. Compare E-Factor to the traditional approach. We will demonstrate how E-Factor equipment
delivers the full package:
POWER LOSSES GRAPH TRANSFORMER OPERATING COST

SAVINGS CALCULATOR
PC2_cover 10/21/04 9:43 AM Page 2

CONNECTING
...PROTECTING
®
®
From overhead to underground...
Hubbell has the products and know-how you need for
planning, constructing and maintaining distribution and
transmission networks. With traditional and innovative
products, Hubbell helps you cut costs, save time and
increase efficiency.
For transmission and distribution, look no further than

Hubbell. We have the products, support and expertise to
ensure your competitive advantage.
Along with the many advantages of
deadfront, Chance AIS Padmounted
Air-Insulated Switchgear also offers a
visible break, fusing and switch configuration
choices, motorized operation, and reduced
long-term maintenance costs.
POWER SYSTEMS 870 Brock Road South • Pickering, ON L1W 1Z8 • Phone (905) 839-1138 • Fax: (905) 831-6353 • www.HubbellPowerSystems.ca
PC2_cover 10/21/04 9:43 AM Page 1

Electrical System Protection and Control Handbook Volume 2

MTE Introduces the Full Spectrum
In motor protection:
Sine Wave Filters and dV/dt Filters
New MTE Sine Wave filters provide
sine wave output voltage when driven
from PWM inverters, eliminating
motor insulation failures, reduce EMI
and help meet IEEE-519 for cable
lengths up to 15,000 feet
New MTE Series A sine-wave filters provide
a sine-wave output voltage when driven from
PWM inverters with switching frequencies from
2 kHz to 8 kHz. For drive applications, these
filters eliminate the problem of motor insulation
failures and reduce electromagnetic interference
by eliminating the high dV/dt associated with
inverter output waveforms in applications
where the distance between the motor and
the inverter is up to 15,000 feet. MTE Sine
Wave Filters permit the use of standard
MTE Sine Wave Filters transformers and meet the requirements of
IEEE—519.
First-turn arc-over protection GUARANTEED for
cable lengths up to 1,000 feet between inverter
and motor with new MTE dV/dt filters
MTE Series A dV/dt filters are designed to protect AC motors
from the destructive effects of peak voltages facilitated by
long cable runs between the inverter and motor. The MTE
dV/dt filter is guaranteed to meet its maximum peak motor
voltage specification (150% of bus voltage) with up to 1,000
Volume 2
feet of cable between the filter and the motor. It is also rated
for a maximum dV/dt of 200 volts per microsecond. In specific
applications, the filter has provided excellent performance
with cable runs up to 3,000 feet. The dV/dt filter has a 3%
insertion impedance which ensures motor torque is not The Electricity Forum
affected by added voltage drops from the filter.
MTE dV/dt Filters
Visit us at: www.mtecorp.com

Or call 1-800-455-4MTE (4683)
THE INTERNATIONAL POWER QUALITY EXPERTS

Protection Control

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Protection Control

Caricato da

Copyright:

Formati disponibili

PC2_cover 10/21/04 9:43 AM Page 1

Buyer: Esam Darweesh (esamdrsh@gmail.com)

Electrical System Protection and Control Handbook Volume 2

Visit us at: www.mtecorp.com

THE INTERNATIONAL POWER QUALITY EXPERTS

Buyer: Esam Darweesh (esamdrsh@gmail.com)

For transmission and distribution, look no further than

Buyer: Esam Darweesh (esamdrsh@gmail.com)

Published by The Electricity Forum

The Electricity Forum The Electricity Forum Inc.

Visit our website at

Buyer: Esam Darweesh (esamdrsh@gmail.com)

2 Electrical System Protection and Control Handbook - Vol.2

The Electricity Forum

© The Electricity Forum 2004

Buyer: Esam Darweesh (esamdrsh@gmail.com)

Electrical System Protection and Control Handbook - Vol.2 3

Testing Consideration for Modern Distance Relays

Asset Management: Purchasing Data Integration for Relay Protection

Protection and Control Upgrade -User Added Value

Analyze Relay Fault Data to Improve Service Reliability

How Certain Features of Protective Devices Help to Protect Workers

Choosing the Right Solid State Relay

Aspects of Overcurrent Protection for Feeders and Motors

Thermally Protected MOV

Rio Grande Electric Monitors Remote Energy Assets Over Satellite

Generator Protection Application Guide

A new approach to high impedance fault detection

BUYERS GUIDE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .93

Buyer: Esam Darweesh (esamdrsh@gmail.com)

4 Electrical System Protection and Control Handbook - Vol.2

Buyer: Esam Darweesh (esamdrsh@gmail.com)

Electrical System Protection and Control Handbook - Vol.2 5

Buyer: Esam Darweesh (esamdrsh@gmail.com)

6 Electrical System Protection and Control Handbook - Vol.2

PROTECTION USING LINEAR COUPLERS

Buyer: Esam Darweesh (esamdrsh@gmail.com)

Electrical System Protection and Control Handbook - Vol.2 7

Algorithms for low-impedance relays are aimed at [8]:

Buyer: Esam Darweesh (esamdrsh@gmail.com)

8 Electrical System Protection and Control Handbook - Vol.2

Fig. 6. Phase-segregated digital busbar scheme.

Buyer: Esam Darweesh (esamdrsh@gmail.com)

Electrical System Protection and Control Handbook - Vol.2 9

Fig.7. Inter-IED communications for sharing on/off signals.

Buyer: Esam Darweesh (esamdrsh@gmail.com)

10 Electrical System Protection and Control Handbook - Vol.2

Fig. 8. Inter-IED communications for sharing on/off signals.

Buyer: Esam Darweesh (esamdrsh@gmail.com)

Electrical System Protection and Control Handbook - Vol.2 11

ADVANTAGES AND BENEFITS

Buyer: Esam Darweesh (esamdrsh@gmail.com)

12 Electrical System Protection and Control Handbook - Vol.2

Buyer: Esam Darweesh (esamdrsh@gmail.com)

Electrical System Protection and Control Handbook - Vol.2 13

Fig. 15. Line-side CT configuration.

Buyer: Esam Darweesh (esamdrsh@gmail.com)

14 Electrical System Protection and Control Handbook - Vol.2

Fig. 16. Bus-side CT configuration.

Fig. 17. Tie-breaker with a single CT.

Buyer: Esam Darweesh (esamdrsh@gmail.com)

Electrical System Protection and Control Handbook - Vol.2 15

COMPLEX SWITCHING STRATEGIES