Paper For Research Topic MS Stduents

Maximizing Line Protection Reliability,
Speed, and Security
Héctor J. Altuve, Karl Zimmerman, and Demetrios Tziouvaras

Schweitzer Engineering Laboratories, Inc.
Presented at the
42nd Annual Western Protective Relay Conference
Spokane, Washington
October 20–22, 2015
1
Maximizing Line Protection Reliability,

Speed, and Security
Héctor J. Altuve, Karl Zimmerman, and Demetrios Tziouvaras
Schweitzer Engineering Laboratories, Inc.
Abstract—This paper describes several commonly applied line requires a different set of tools, such as computer-based power
protection schemes, including distance schemes, directional system studies and event information analysis.
comparison schemes using distance and directional elements, and This paper briefly describes several commonly applied line
line current differential schemes. Using analysis tools like fault
trees, power system studies, and event analysis, we evaluate and
protection schemes. Using analysis tools like fault trees,
compare these protection schemes in terms of speed, sensitivity, power system studies, and event analysis, we evaluate these
dependability, security, and selectivity. The paper considers the schemes in terms of speed, sensitivity, dependability, security,
use of various communications channels, including direct relay- and selectivity and provide a comparison of the schemes’
to-relay fiber-optic channels and multiplexed digital fiber-optic performance. The paper mainly considers the use of fiber-
networks. The paper also discusses some practical considerations optic communications channels, but it also provides data on
for evaluating line protection schemes when faced with
complications like series compensation, mutual coupling, single-
the speed of other channels, such as power line carrier (PLC)
pole tripping and reclosing, three-terminal lines, and short lines. and digital radios. The paper also discusses some practical
considerations for evaluating line protection schemes in
I. INTRODUCTION complex applications.
Traditional protection systems consisted of a number of
single-function electromechanical relays that provided good II. BASIC CONCEPTS
service for many years. However, providing redundancy A. Protection System Functional Characteristics
required duplicating many devices. In addition, increased Protection system functional characteristics must meet the
maintenance costs, lack of support by manufacturers, and stringent requirements of modern power systems, which lack
incorrect operation data led many utilities to replace redundancy and operate near their security limits. The most
electromechanical relays with microprocessor-based relays, important characteristics are reliability, selectivity, speed of
which provide better protection and control functions at lower operation, and sensitivity.
cost and with higher reliability. They also have monitoring • Reliability is a measure of the certainty that the
and communications abilities. Redundant protection systems protection system will trip when required
are more economical with multifunction microprocessor-based (dependability) and not trip when not required
relays. On the other hand, microprocessor-based relays have (security). We can obtain dependability through relays
shifted complexity from panel designs and wiring to settings, that try to trip the same circuit breaker (parallel
logic, and documentation. connection of the relay contacts or its equivalent logic
Modern power systems demand that transmission line function [OR logic]). We can obtain security through
protection schemes be reliable (dependable and secure), fast, series connection of the relay contacts or the
sensitive, and selective. However, these protection system equivalent logic function (AND logic). There is a bias
characteristics are frequently at odds with each other. For among protection engineers toward dependability in
example, in a dual-redundant system, we need to connect the protection system design. This bias reflects the fact
relay output contacts in parallel to achieve dependability but that power systems are redundant to a certain extent.
connect them in series to achieve security. A very fast and In modern power systems, however, this concept is
highly sensitive protection system may not be very secure or changing. For example, some wide-area protection
selective. For this reason, engineers have traditionally needed systems, where security is very important, use two-
to make design choices to prioritize some protection system out-of-three voting schemes.
characteristics at the expense of others. With today’s • Selectivity is the ability of a protection system to
technology, is it possible to simultaneously maximize all of eliminate a fault in the shortest time possible with the
these critical protection characteristics? least disconnection of system components. We also
References [1] and [2] discuss the application of fault tree use the term coordination for selectivity. Protection
analysis to determine factors that influence overall protection coordination implies that primary protection
system reliability and provide comprehensive data of eliminates faults and that backup protection operates
reliability indices. In particular, [1] evaluates line protection only when primary protection fails.
redundancy and reliability. Analyzing protection system
speed, sensitivity, and selectivity not covered in [1] and [2]
2
• Speed of operation is the ability of the protection Availability and unavailability are dimensionless numbers
system to operate in a short time after fault inception. from 0 to 1. We can convert them to minutes or seconds per
Fast operation is important in preserving system year by multiplying by the appropriate factors.
stability, reducing equipment damage, and improving Fault tree analysis is a tool for evaluating how a component
power quality. Protection system operation time failure contributes to a specific failure event [3] [4]. Fault tree
includes relay operating time, communications system analysis is useful for comparing the relative reliability of
delay (if any), and circuit breaker fault-clearing time. protection schemes. The failure event of interest is called the
• Sensitivity is the ability of the protection system to top event. The failure rate for the top event is a combination of
detect even the smallest faults within the protected the failure rates of the basic events that contribute to the top
zone. It is important to ensure the detection of high- event. Basic events are individual component failures with
impedance faults or the reduced contribution to faults identified failure rates. We use AND, OR, and other gates to
from small, distributed generators. represent combinations of failure rates. OR gates express the
idea that any of several failures can cause the protection
B. Basic Reliability Concepts system to fail. The OR gate output is the sum of the failure
We often use the following measures to describe product rates of the input events. AND gates express the idea that
reliability performance, assuming constant failure and repair failures must occur simultaneously to cause the protection
rates [1] [2] [3]: system to fail. The AND gate output is the product of the
• Failure: Termination of the ability of an item to failure rates of the input events. We can also use availability,
perform its required or specified function. unavailability, or MTBF figures instead of failure rates in fault
• Failure rate (λ): Total number of failures divided by tree analysis.
the total unit operating time or uptime. Analyzing the dependability and security of a protection
• Repair rate (µ): Total number of repairs divided by the system requires different fault trees [1] [2]. To construct each
total unit operating time or uptime. tree, we identify which component failures may cause a failure
• Mean time to failure (MTTF): Average time between to trip (a dependability problem) or an undesired trip (a
the start of operation (or return after repair) and security problem). This analysis leads to different tree
failure. For a constant failure rate, MTTF = λ–1. topologies and different failure rate (or unavailability) values.
• Mean time to repair (MTTR): Average time to correct For example, nearly any relay failure could cause a failure to
a failure and restore a unit to operating condition. For trip. However, not all relay failures cause an undesired trip.
a constant repair rate, MTTR = µ–1. Hence, the relay failure rate or unavailability value to use for
• Mean time between failures (MTBF): Average time dependability analysis is higher than the value to use for
between failures for units repaired and returned to use. security analysis. In this paper, we use unavailability for
MTBF is the sum of MTTF and MTTR. Because MTTR is dependability fault trees because failures to clear faults depend
usually small compared to MTTF, we assume that MTBF is on component downtime per failure. We use failure rate for
approximately equal to MTTF and that MTBF = λ–1. security fault trees because undesired trips typically occur at
Availability, a measure that considers repeated cycles of the instant a component fails [5].
failure and repair, is the probability or fraction of time that a
device or system is able to operate. Equation (1) defines III. EXAMPLE SYSTEMS
availability A for constant failure and repair rates. A. Example Power Systems
µ MTTF MTBF We performed computer-based steady-state power system
=A = ≈ (1)
λ + µ MTTF + MTTR MTBF + MTTR studies to evaluate the sensitivity, speed, and selectivity of
Unavailability is the probability or fraction of time a device various line protection schemes. We selected three two-source
or system is unable to perform its intended function. example systems with two parallel transmission lines with
Equation (2) defines unavailability U for constant failure and magnetic mutual coupling. Fig. 1(a) shows a system with two
repair rates. strong sources and two long lines. Fig. 1(b) depicts a system
with two strong sources and two short lines. Fig. 1(c) shows a
λ MTTR system with a strong source, a weak source, and two long
U 1 –=
= A = ≈ λ MTTR (2)
λ + µ MTBF lines.
3
Bus S Bus R B. Example Protection Systems

[m = 0] [m = 1]
Reference [1] describes a reliability study of several
C ZL2 D directional comparison line protection schemes using fault
ZS Z0M ZR
trees. This study covered permissive overreaching transfer trip
(POTT) and directional comparison blocking (DCB) schemes
Source S A ZL1 B Source R with PLC, microwave, and fiber-optic communications
channels. The study evaluated the effect of protection system
R R redundancy; comprehensive commissioning testing; using
relays from the same or different manufacturers; and common-
Z1S = 2∠88° Ω Z1L1 = Z1L2 = 8∠84° Ω Z1R = 2∠88° Ω mode failures in relays. Table V of the Appendix summarizes
Z0S = 2∠88° Ω Z0L1= Z0L2= 24∠80° Ω Z0R = 2∠88° Ω the results.
Z0M= 16∠78° Ω In this paper, we extend the study in [1] to include line
(a)
current differential (87L) schemes. We compare the reliability
of POTT and 87L schemes protecting a transmission line with
Bus S Bus R
[m = 0] [m = 1] single circuit breakers at both ends. All schemes use
multifunction relays and fiber-optic communications channels.
C ZL2 D
Fig. 2(a) shows the basic line protection scheme, which
ZS Z0M ZR consists of a communications-based protection scheme
(Relay R1) complemented with a separate distance protection
Source S A ZL1 B Source R scheme (Relay R2) at each line end. The scheme includes one
set of instrument transformers, one dc power system, and a
R R circuit breaker with a single trip coil at each line end. In this
scheme, the Zone 1 elements of the distance protection
Z1S = 2∠88° Ω Z1L1
= Z1L2
= 0.4∠84° Ω Z1R = 2∠88° Ω scheme provide redundant protection, independent from the
Z0S = 2∠88° Ω Z0L1
= Z0L2= 1.2∠80° Ω Z0R = 2∠88° Ω communications channel. The scheme lacks high-speed
Z0M
= 0.8∠78° Ω
tripping redundancy for faults that fall outside the reach of
(b) Zone 1 elements.
Bus S Bus R
Fig. 2(b) shows a dual-redundant protection scheme, which
[m = 0] [m = 1] includes two communications-based schemes with separate
fiber-optic communications channels, two relays, two sets of
C ZL2 D
instrument transformers, two dc power systems, and a circuit
ZS Z0M ZR breaker with redundant trip coils at each line end. To create a
A ZL1 B
triple-redundant scheme, we add a third communications-
Source S Source R
based scheme to the Fig. 2(b) scheme. In redundant schemes,
we assume that all of the redundant components are of similar
R R
quality and that relays have the same reliability indices.
Fig. 2 represents the communications channel as a digital
Z1S = 2∠88° Ω Z1L1 = Z1L2 = 8∠84° Ω Z1R= 40∠60° Ω multiplexed fiber-optic network. We can replace the
Z0S = 2∠88° Ω Z0L1= Z0L2= 24∠80° Ω Z0R= 40∠60° Ω
multiplexers and the network with an optical fiber providing
Z0M= 16∠78° Ω
direct relay-to-relay communication.
(c)
Fig. 1. Example power systems: (a) two long lines and strong sources; (b)
two short lines and strong sources; and (c) a strong source at S, a weak source
at R, and two long lines.
4
Bus S Bus R
52 52
R2 R2
Fiber-Optic Digital Fiber- Fiber-Optic

R1 R1
Multiplexer Optic Network Multiplexer
125 Vdc 48 Vdc 48 Vdc 125 Vdc

(a)
125 Vdc 2 48 Vdc 2 48 Vdc 2 125 Vdc 2

R2 R2
Bus S Bus R
52 52

R1 R1
125 Vdc 1 48 Vdc 1 48 Vdc 1 125 Vdc 1

(b)
Fig. 2. Single- and dual-redundant transmission line protection schemes. An optical fiber providing direct relay-to-relay communication can replace the
multiplexer and digital fiber-optic network.
protection algorithms: directional overcurrent, distance, and

IV. PROTECTION SCHEME SPEED ANALYSIS
current differential. For example, Fig. 3 shows the functional
Power system stability continues to drive the quest for block diagram of a typical microprocessor-based mho relay
faster protection. Faults must be cleared faster than critical element. The analog low-pass filters reject the high-frequency
clearing times or systems may lose transient stability. Faster signal components to avoid aliasing errors in the sampling
protection also allows increased power transfer capability, process. The digital band-pass filters extract the fundamental
reduces equipment damage, and improves power quality. frequency components of the sampled and digitized voltage
The fault-clearing time has the following components: and current signals. In Fig. 3, VRE, VIM, IRE, and IIM designate
• Protection scheme tripping time (PSTT): The time the real and imaginary components of the voltage and current
elapsed between fault inception and the instant when phasors, respectively. The relay performs a phase comparison
the protection scheme issues a circuit breaker tripping between an operating signal derived from the voltage and
signal. current phasors and a polarizing signal (typically the
• Circuit breaker fault-clearing time. memorized positive-sequence voltage) to create the mho
PSTT consists of the sum of the delays of all devices that characteristic shown in Fig. 3. Directional overcurrent and
must operate in order for the protection scheme to produce a current differential elements also process phasors obtained as
circuit breaker tripping signal. PSTT includes: in Fig. 3.
• Relay operating time: Includes the protective relay and
Analog Digital VRE
auxiliary relay (if used) delays. v Low-Pass Band-Pass
X
VIM
• Communications system delay (for communications- Filter Analog/ Filter
based protection schemes). Digital
Analog Conversion Digital IRE R
i
A. Protective Relay Operating Time Low-Pass Band-Pass
IIM
Filter Filter
Today’s line protection schemes typically use
microprocessor-based relay elements with phasor-based Fig. 3. Functional block diagram of a phasor-based mho element.
5
Fig. 4 shows the different components of the operating principle allows reducing the relay operating time to a quarter
time of a microprocessor-based relay with phasor-based of a cycle or less [9].
protection algorithms. Analog low-pass filter delay and
One-Cycle One-Cycle Mho
sampling latency determine how fast the signal samples are v, i Filter Calculation –
available for processing by the digital filter. The analog filter Zone 1 +
delay depends on the filter type and its cutoff frequency, Reach
which depends on the sampling rate. For a 16 samples per Zone 1
Detection
cycle rate, the analog filter delay is around 0.04 cycles. After Half-Cycle Half-Cycle Mho
Filter Calculation –
the delay introduced by the analog filter, the digital band-pass
filter still needs to wait for the next sample to be available for Reduced +
Zone 1
processing. This is the sampling latency, whose value is Reach
between zero (when the fault occurs just before the next
Fig. 5. Zone 1 mho distance element using dual-filter scheme.
sampling instant) and the sampling period (for faults occurring
just after the last sampling instant). For a 16 samples per cycle The relay output system produces the contact closing
rate, the sampling period is 0.063 cycles. The digital band- operation that completes the circuit breaker tripping coil
pass filter typically introduces the longest delay. Reducing this circuit. Conventional output systems with electromechanical
delay is instrumental to achieving high-speed relay operation. relays typically add a 4 to 6 ms delay. We can use high-speed
The digital filter delay is determined by the data window output systems with solid-state circuits that operate in about
length and the input signal magnitude as compared to the 10 µs to substitute conventional output systems.
pickup setting. The digital filter delay approaches the data Fig. 6 and Fig. 7 depict the typical operating times of two
window length for faults representing an operation condition types of phase and ground mho elements, one that uses the
close to the relay pickup setting. For higher fault currents, the Fig. 5 dual-filter scheme and another that uses only one-cycle
digital filter delay is smaller than the data window length. filters. These elements have high-speed output systems;
Relay protection algorithms process the phasors estimated by adding 4 to 6 ms to these times gives the typical operating
the digital filter to make tripping decisions. This processing times for mho elements with conventional output systems.
introduces a delay, as shown in Fig. 4. Finally, the relay Fig. 6 and Fig. 7 show the operating time as a function of fault
output system requires time to process the tripping signal and distance (in percent of reach setting) for different values of
close the output relay contact. This is the output device delay. source-impedance ratio (SIR). Fig. 6 and Fig. 7 show that the
high-speed mho elements consistently trip in less than one
Analog Filter Delay Output Device Delay
cycle.
Sampling Latency Protection Algorithm
Processing Delay
High-Speed Elements
Digital Filter Delay
1.50
1.25
Time (Cycles)
SIR = 10.0
1.0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 0.75 SIR = 1.0
Time (Cycles) 0.50
Fig. 4. Components of the operating time of a microprocessor-based relay 0.25 SIR = 0.1
with phasor-based protection algorithms.
20 40 60 80
As mentioned before, filtering delays relay operation. Most Fault Location (Percent of Reach Setting)
of this delay comes from the digital band-pass filter. The
choice of the digital filter data window length (half cycle, one Standard-Speed Elements
cycle, and so on) directly impacts relay operating speed and

1.50
transient performance [6]. The longer the filter window, the
1.25 SIR = 10.0
longer the delay, but the lower the transient overreach. The
Time (Cycles)
1.0
shorter the window, the shorter the delay, but the greater the
SIR = 1.0
transient overreach. In order to maximize the performance, we 0.75
can combine long- and short-window filters to achieve faster 0.50

SIR = 0.1
speeds for close-in faults and good transient performance for 0.25
zone-boundary faults. Fig. 5 shows a dual-filter scheme that
20 40 60 80
uses a full-cycle mho element in parallel with a high-speed,
Fault Location (Percent of Reach Setting)
half-cycle element. In order to ensure high-speed element
security, the relay reduces the reach of the high-speed element Fig. 6. Typical operating time of phase mho elements with high-speed, dual-
to compensate for the increased transient overreach [7] [8]. A filter schemes and with standard-speed, one-cycle filters.
better approach to improve speed is to apply time-domain
protection principles. For example, the traveling wave
6
High-Speed Elements B. Pilot Protection Schemes

Pilot protection uses a communications channel to compare
1.50
information from the line terminals and provide high-speed
1.25
Time (Cycles)
SIR = 10.0 fault clearing for 100 percent of the protected line. Pilot
1.0
protection includes directional comparison schemes and
0.75 SIR = 1.0 current-based schemes (phase comparison and 87L schemes).
0.50 In a directional comparison scheme, instantaneous
0.25 SIR = 0.1 directional overcurrent or distance elements provide fault
direction information for the scheme logic at each line
20 40 60 80
terminal. Directional comparison does not require a high-
bandwidth channel because the relays exchange information
Standard-Speed Elements on the status of their directional or distance elements. Typical
bandwidth requirements are 0.5 to 1.5 kHz for analog
1.50 channels and 9.6 kbps for digital channels. Directional
1.25 SIR = 10.0 comparison schemes include:
Time (Cycles)
1.0 • Direct underreaching transfer trip (DUTT).

0.75 SIR = 1.0 • Permissive underreaching transfer trip (PUTT).
0.50 • Permissive overreaching transfer trip (POTT).
SIR = 0.1
0.25 • Directional comparison blocking (DCB).
• Directional comparison unblocking (DCUB).
20 40 60 80
Microprocessor-based 87L schemes perform the
differential comparison of the protected line terminal currents.
Fig. 7. Typical operating time of ground mho elements with high-speed, The relays can exchange digitized current samples or current
dual-filter schemes and with standard-speed, one-cycle filters. phasor values. 87L protection requires a digital microwave or
Microprocessor-based 87L schemes perform the fiber-optic channel with a bandwidth of 56 kbps or higher.
differential comparison of digitized current samples or current The communications system delay causes a fictitious phase
phasor values from all line terminals. Fig. 8 depicts the typical shift between the local current and the received remote
operating time of phase (87LP), negative-sequence (87LQ), current(s). 87L schemes align (synchronize) current samples
and zero-sequence (87LG) differential elements. The 87LP or phasors to prevent the errors caused by this phase shift.
elements trip in less than one cycle for differential currents Typical communications system delay requirements for 87L
above three times pickup current. schemes are in the range of 5 to 10 ms.
1.6
C. Components of PSTT in Pilot Protection Schemes
The pilot protection schemes of all line terminals must
1.4
operate to achieve high-speed fault clearing. Hence, PSTT is
87LQ the time it takes the slowest scheme to issue the circuit breaker
Time (Cycles)
1.2
87LG tripping signal.
1 Fig. 9(a) shows the components of PSTT in a traditional
87LP
directional comparison tripping scheme (PUTT, POTT, or
0.8
DCUB) with analog relays and an analog PLC or microwave
0.6
channel. PSTT includes the local relay operating time, local
1.4 2 3 4 5 7.5 10 19 and remote communications equipment delays,
Per Unit Differential Current
communications channel delay, processing delay at the remote
Fig. 8. Typical operating time of phase (87LP), negative-sequence (87LQ), relay, and remote relay output device delay. Microprocessor-
and zero-sequence (87LG) differential elements. based relays start the signal transmission process when the
The PSTT of the 87L scheme includes the differential output of a protection element asserts, as shown in Fig. 9(b).
element operating time and the communications system delay There is no need to wait for the relay output contact to close.
(the time it takes for the remote end current information to Direct relay-to-relay communication over optical fiber
reach the local differential element). Section D covers practically eliminates the communications equipment delay
communications system delay. (the fiber-optic transceiver is very fast), as shown in Fig. 9(c).
7
Communications Communications latency. Modern communications equipment uses high-speed,

Protection Element Equipment Equipment Output
Operating Time Delay Delay Device Delay solid-state outputs. However, some legacy systems require
interposing relays that add 2 to 4 ms of delay and are also
Relay
Output Channel Processing susceptible to contact bounce, for which an additional delay
Device Delay Delay Delay may be necessary to “debounce” inputs [10].
As an example, Table I, taken from [11], lists typical
processing delays for relays using a proprietary relay-to-relay
communications protocol for different processing rates and
Relay Operating Time
Communications channel speeds.
System Delay
TABLE I
TYPICAL PROCESSING DELAYS FOR RELAYS USING A RELAY-TO-RELAY
PSTT COMMUNICATIONS PROTOCOL [11]
(a) Relay Processing Delay (ms)

Channel
Speed (kbps) 8 Samples per 4 Samples per
Protection Element Channel Relay Processing
Operating Time Delay Delay
Cycle Cycle
Communications Communications 38.4 4.2 8.3

Equipment Equipment Output
Delay Delay Device Delay 19.2 6.3 10.5
9.6 8.3 12.5
4.8 12.5 16.7
Communications Fig. 10 shows the typical time chart for a POTT scheme
System Delay
with overreaching distance elements, excluding Zone 1 direct
tripping. DCB schemes require an additional coordinating
PSTT
time delay to wait for the blocking signal to arrive from the
(b) remote terminal(s). For this reason, DCB schemes are slightly
slower than PUTT, POTT, or DCUB schemes.
Protection Element Relay Processing
Operating Time Delay Local POTT
Fault Scheme Trip
Channel
Output Local
Delay
Device Delay
Overreaching
Communications,
Distance
Processing, and
Element
Output Device Delay
Received
Permissive
Trip
Remote POTT
Remote
Scheme Trip
PSTT Overreaching
Distance
Element
(c)
Received
Fig. 9. Components of PSTT in line pilot protection schemes: (a) directional Permissive
comparison tripping scheme (PUTT, POTT, or DCUB) with analog relays and Trip Communications,
analog PLC or microwave channel; (b) pilot protection scheme with Processing, and
microprocessor-based relays and digital microwave channel, digital radio Output Device Delay
channel, or digital fiber-optic network; and (c) pilot protection scheme with
direct relay-to-relay communication over optical fiber. Both POTT
PSTT Schemes Trip
The relay processing delay in Fig. 9 includes the hardware
and firmware delays (approximately 3 ms), plus the Fig. 10. Time chart for a POTT scheme with overreaching distance elements
processing latency. This latency equals zero when an input is (excluding Zone 1 direct tripping).
processed just before a change of state, and it equals the D. Communications System Delay
processing period when the input is processed just after a
Fig. 9(a) and Fig. 9(b) show that the communications
change of state. The longest processing latency is 4 ms for a
system delay (also called communications system latency)
4 sample/cycle processing rate and 2 ms for an 8 sample/cycle
includes the equipment and channel delays.
rate. A slow channel further increases relay processing
8
PLC systems, typically used for DCB schemes, transmit 900 MHz range and are limited to line of sight. Under
information over the line conductors. PLC frequencies are favorable conditions, the radios can work on lines up to about
usually in the 30 to 500 kHz range. Table II shows typical 20 miles. Table IV shows typical communications delays for
delays for traditional on-off and frequency-shift PLC systems. digital relay-to-relay communications systems using point-to-
Always check manufacturer published specifications when point radios. The use of AES-256 encryption introduces some
determining operating times for any given scheme. Wide-band additional delay.
PLC systems are faster than narrow-band systems, but the TABLE IV
wider band channel allows more noise into the receiver filter, TYPICAL DELAYS FOR DIGITAL RELAY-TO-RELAY
which introduces a greater chance for a false assertion. COMMUNICATIONS SYSTEMS USING POINT-TO-POINT RADIOS
TABLE II Delay (ms)

TYPICAL DELAYS FOR TRADITIONAL PLC SYSTEMS Channel Speed
(kbps) With AES-256
No Encryption
Communications Equipment Delay (ms) Encryption
PLC (wide or narrow band) 4 to 8 [10] 38.4 4.8 N/A
PLC (on/off DCB) 2 to 4 19.2 5.6 7.4
Pilot protection schemes using digital communications E. Evaluation of PSTT by Computer Simulations
technology have been in use for over 20 years. These schemes In order to evaluate the operating speed of different
use direct or multiplexed optical fiber, digital radios, and protection schemes, we simulated faults at different locations
digital microwave channels. on one line of each of the three example systems shown in
Considering a direct relay-to-relay connection over optical Fig. 1. We determined the average PSTT values for POTT and
fiber, the communications delay depends on the time required 87L schemes. We considered the following pilot schemes:
for light to travel over the optical fiber (0.8 ms per 100 miles). • POTT scheme with distance Zone 1 direct tripping:
In multiplexed fiber-optic networks, multiplexers and − POTT scheme using Zone 1 and Zone 2 phase and
repeaters introduce additional delays. For protection ground high-speed distance elements (the Zone 1
multiplexers, the delay is only 0.5 ms plus approximately element also provides direct circuit breaker
24 µs for each repeater in the path. Some multiplexers tripping).
designed for telecommunications applications over Ethernet
− Zone 1 reach set to 80 percent of the line length for
do not perform to protection expectations. These multiplexers long lines and 60 percent for short lines.
may introduce 6 to 8 ms delays. Table III shows typical delays
− Zone 2 reach set to 200 percent of the line length.
for direct relay-to-relay communication over optical fiber and
− Communication at 38.4 kbps over fiber-optic
multiplexed fiber-optic networks.
network with multiplexers (for long lines).
TABLE III − Relay-to-relay communication at 38.4 kbps over
TYPICAL DELAYS FOR DIRECT RELAY-TO-RELAY COMMUNICATION OVER
OPTICAL FIBER AND MULTIPLEXED FIBER-OPTIC NETWORKS direct optical fiber (for short lines).
− Fast (<10 µs) output devices.
Communications System
Delay • 87L scheme with distance Zone 1 direct tripping:
Component
− 87L scheme using phase (87LP), negative-
Optical fiber 0.8 ms per 100 miles sequence (87LQ), and zero-sequence (87LG)
Protection-class multiplexer 0.5 ms elements. 87LP pickup current set to 1.2 per unit.
Nonprotection-class multiplexer 6 to 8 ms
87LQ and 87LG pickup current set to 0.2 per unit.
− Zone 1 reach set to 80 percent of the line length for
Repeater 24 µs/repeater
long lines and 60 percent for short lines.
For example, for direct relay-to-relay communication − Communication at 56 kbps over fiber-optic
systems, the delay is nearly zero for a short line since the network with multiplexers (for long lines).
optical fiber delay is less than 0.8 ms. For multiplexed fiber- − Relay-to-relay communication at 56 kbps over
optic networks, the communications system delay depends on direct optical fiber (for short lines).
the line length and the fiber-optic network delay. Using − Fast (<10 µs) output devices.
protection multiplexers, the communications system delay is Fig. 11 through Fig. 16 show the average PSTT values
typically under 2 ms. For a 100-mile line, the delay would be resulting from computer simulations.
about 1.3 ms, based on 0.8 ms (optical fiber delay), plus
approximately 0.5 ms (multiplexer delay), plus approximately
24 µs per repeater [12].
In some applications, such as subtransmission line
protection, where fault-clearing speed is not as critical, digital
point-to-point serial radios are an acceptable and economical
alternative. These radios typically operate in the unlicensed
9
1.5 1.5
PSTT (Cycles)
PSTT (Cycles)
Phase-to- Phase-to-
Ground Fault Ground Fault
1.0 1.0
Phase-to- Phase-to-
Phase Fault Phase Fault
0.5 0.5
0 0.25 0.50 0.75 1.0 0 0.25 0.50 0.75 1.0

m (Per Unit Distance to Fault) m (Per Unit Distance to Fault)
Fig. 11. Average PSTT values for POTT scheme and 80 percent distance
Zone 1 direct tripping. System with two long lines and strong sources.
Zone 1 direct tripping. System with two long lines, a strong source at the local
end, and a weak source at the remote end.
1.5
PSTT (Cycles)
Phase-to- 1.5
Ground Fault Phase-to-
PSTT (Cycles)
Ground Fault
1.0
Phase-to- 1.0
Phase-to-
Phase Fault Phase Fault
0.5
0.5
0 0.25 0.50 0.75 1.0

m (Per Unit Distance to Fault) 0 0.25 0.50 0.75 1.0
m (Per Unit Distance to Fault)
Fig. 12. Average PSTT values for 87L scheme and 80 percent distance
Zone 1 direct tripping. System with two long lines and strong sources. Fig. 16. Average PSTT values for 87L scheme and 80 percent distance
Zone 1 direct tripping. System with two long lines, a strong source at the local
end, and a weak source at the remote end.
1.5
Fig. 11 through Fig. 16 show that modern protection
schemes using high-speed elements and fast communications
PSTT (Cycles)
Phase-to-Phase or
Phase-to-Ground Fault channels produce low and consistent PSTT values for various
1.0 line lengths and source strengths.
F. Actual Fault Case
0.5
Many power system events demonstrate the speed of
modern directional comparison and 87L protection schemes.
For example, [13] reports multiple relay operations that show
0 0.25 0.50 0.75 1.0
m (Per Unit Distance to Fault) PSTT values close to one cycle, including one event in which
the 87L scheme detects two consecutive single-phase-to-
Zone 1 direct tripping. System with two short lines and strong sources.
ground faults in 0.75 cycles, as shown in Fig. 17.
1.5
PSTT (Cycles)
Phase-to-Phase or
Phase-to-Ground Fault
1.0
0.5
0 0.25 0.50 0.75 1.0

m (Per Unit Distance to Fault)
Fig. 14. Average PSTT values for 87L scheme and 60 percent distance
Zone 1 direct tripping. System with two short lines and strong sources.
10
5000
87LG, 87LQ
Current
4000
3000 125
2000
1000
0
67N
–1000
–2000
67N
–3000 100 67Q
–4000
–5000
0.75 cycles 0.75 cycles
67Q
RF (Ohms Secondary)
14.775950 14.875950 14.975950 15.075950 15.175950 15.275950
600
400
75
Voltage
200
–200
50
–400
–600
14.775950 14.875950 14.975950 15.075950 15.175950 15.275950
21X 21X
1:87LG
25
1:87LQ
1:TRIP 21N 21N

1:87LP
0
0 0.5 1
14.775950 14.875950 14.975950 15.075950 15.175950 15.275950
m (Per Unit of Line Length)
Fig. 17. 87L scheme detects two consecutive faults in 0.75 cycles. Fig. 18. RF coverage of 21N, 21X, 67N, 67Q, 87LG, and 87LQ elements.
System with two long lines and strong sources.
V. PROTECTION SYSTEM SENSITIVITY ANALYSIS
87LG, 87LQ
125
Measuring fault resistance (RF) coverage is an excellent
way to evaluate the sensitivity of a protection system.
References [14], [15], and [16] evaluate the sensitivity of 100
RF (Ohms Secondary)
ground directional and distance elements and ground pilot
protection schemes based on RF coverage. In this paper, we 75 67N 67N
67Q 67Q
expand this discussion by including an evaluation of 87L
element sensitivity and comparing it to distance and 50
directional element sensitivity.
A. Protection Element RF Coverage 25
In order to evaluate sensitivity, we simulated single-phase-
21N, 21X 21N, 21X
to-ground faults at different locations on one line of each of 0
0 0.5 1
the three example systems shown in Fig. 1. We determined the m (Per Unit of Line Length)
maximum value of RF detected by each protection element for
Fig. 19. RF coverage of 21N, 21X, 67N, 67Q, 87LG, and 87LQ elements.
each fault location. We used the following settings for the
System with two short lines and strong sources.
studied protection elements:
87LG, 87LQ
• Ground mho distance element (21N): Reach setting of 125
2Z1L. 67Q
• Ground quadrilateral distance element (21X): 100 67N 67Q
− Reactance reach setting of 2Z1L.
RF (Ohms Secondary)
− Resistance reach setting of 50 ohms secondary.

75
• Zero-sequence directional overcurrent element (67N):
Pickup current setting of 3I0 = 0.5 A. 67N
21X
• Negative-sequence directional overcurrent element 50
(67Q): Pickup current setting of 3I2 = 0.5 A. 21X
• Zero-sequence differential element (87LG): 25
Differential current pickup setting of 3I0 = 0.5 A. 21N 21N
• Negative-sequence differential element (87LQ): 0
Differential current pickup setting of 3I2 = 0.5 A. 0 0.5 1
Fig. 18 through Fig. 20 depict the RF coverage of 21N,
21X, 67N, 67Q, 87LG, and 87LQ elements as a function of Fig. 20. RF coverage of 21N, 21X, 67N, 67Q, 87LG, and 87LQ elements.
fault distance. These figures show that the value of RF System with two long lines, a strong source at the local end, and a weak
source at the remote end.
detected by the ground distance and directional elements
decreases as the fault moves away from the relay location B. Combining Speed and Sensitivity
because these elements measure only the local current, which Reference [16] compares various pilot protection schemes,
diminishes as the fault moves away. However, 87LG and including an evaluation of speed and sensitivity together. To
87LQ elements measure the total fault current (the sum of the add to this work, we also considered 87L schemes. For which
local and remote currents), which maximizes their RF faults can we maximize both speed and sensitivity?
coverage and makes it independent of the fault location.
11
Fig. 21 shows the RF coverage regions of a POTT scheme 125

87LG, 87LQ
using instantaneous overreaching 67N elements for the system
with two long lines and strong sources shown in Fig. 1(a).
Fig. 21 does not show the effect of the directly tripping 100
underreaching instantaneous 67N elements or the time-
RF (Ohms Secondary)
delayed overreaching 67N elements. The POTT scheme trips
75
in less than 1.5 cycles when the overreaching 67N elements of 67N, 67Q
both line terminals detect the ground fault (Region A in
Fig. 21). When only one overreaching 67N element detects the
50
fault, the POTT scheme does not trip immediately because it
does not receive the permissive tripping signal from the other
21X
terminal. If an instantaneous underreaching 67N element trips 25
the circuit breaker, the fault current redistribution with that
terminal open allows the overreaching 67N element of the
other terminal to detect the fault and send the permissive 0
0 0.5 21N 1
signal. This sequential operation introduces some fault-
clearing delay (Region B in Fig. 21). If no underreaching 67N
element detects the fault, no current redistribution occurs, and Fig. 22. High-speed RF coverage regions of POTT schemes with 21N, 21X,
67N, or 67Q elements and of 87L schemes with 87LG or 87LQ elements.
the POTT scheme does not operate. Fault clearing occurs System with two long lines and strong sources.
when the time-delayed overreaching 67N elements trip the
87LG, 87LQ
circuit breakers. This time-delayed fault clearing defeats the 125
POTT scheme purpose.
RF Coverage With Remote Terminal Open
100
125
RF (Ohms Secondary)
Tripping Delayed by
Sequential Operation 75
67N, 67Q
100
RF (Ohms Secondary)
50
75
Region B Region B
25
50 21N, 21X
Region A
0
0 0.5 1
25
High-Speed Tripping at
Both Ends Fig. 23. High-speed RF coverage regions of POTT schemes with 21N, 21X,
0 67N, or 67Q elements and of 87L schemes with 87LG or 87LQ elements.
0 0.5 1 System with two short lines and strong sources.
87LG, 87LQ
125
Fig. 21. RF coverage regions of a POTT scheme using instantaneous
overreaching 67N elements for the system with two long lines and strong
sources.
100
Fig. 22 through Fig. 24 show the high-speed RF coverage
RF (Ohms Secondary)
regions (corresponding to Region A in Fig. 21) for the 67Q

75
different line protection schemes studied in this paper and for
the different power systems of Fig. 1. These figures show the 67N
higher sensitivity and speed of 87L schemes using 87LG and 50
87LQ elements as compared with POTT schemes using either
67N or 21N (or 21X) elements. Fig. 22 through Fig. 24 also 21X
25
show that 67N elements provide higher sensitivity and speed
than 21N (or 21X) elements.
0
0 0.5 1
21N
Fig. 24. High-speed RF coverage regions of POTT schemes with 21N, 21X,
67N, or 67Q elements and of 87L schemes with 87LG or 87LQ elements.
System with two long lines, a strong source at the local end, and a weak
source at the remote end.
12
C. Actual Fault Case • Dual-redundant POTT/21 and 87L/21 scheme

One actual system fault on a 525 kV transmission line with [Fig. 2(b)]. One relay performs POTT and 21 scheme
500 ohms primary (or 44 ohms secondary) of fault resistance functions and the other performs 87L and 21 scheme
was cleared by a pilot protection scheme using 67N elements functions.
[17]. Fig. 25 shows the phasors measured by the relay of one • Dual-redundant POTT and 87L scheme [Fig. 2(b)].
line terminal. If an 87L scheme had been applied on this line, One relay performs POTT scheme functions and the
the relays at both terminals would have measured other performs 87L scheme functions (no direct
approximately 190 A of negative-sequence and zero-sequence Zone 1 tripping).
differential current. These primary current values yield 3I0 = • Dual-redundant 87L/21 scheme [Fig. 2(b)]. Each relay
3I2 = 1.425 A secondary, well above the minimum pickup performs 87L and 21 scheme functions.
current of modern 87LG and 87LQ elements. • Dual-redundant 87L scheme [Fig. 2(b)]. Each relay
performs 87L scheme functions (no direct Zone 1
tripping).
• Triple-redundant two-out-of-three voting 87L/21
scheme. Each relay performs 87L and 21 scheme
functions.
• Triple-redundant two-out-of-three voting 87L scheme.
Each relay performs 87L scheme functions (no direct
Zone 1 tripping).
Table VI of the Appendix shows the reliability indices used
in the fault trees.
In this section, we describe several fault trees and
summarize the results obtained from all of the fault trees.
A. Single Schemes
Fig. 26 shows the dependability fault tree for the
Fig. 25. Currents and voltages measured by the relay during a ground fault combination of POTT/21 and 21 schemes in separate relays
on a 525 kV line with RF = 500 ohm primary.
[Fig. 2(a)] with a fiber-optic channel. The top event is
“protection fails to clear in-section fault in the prescribed
VI. PROTECTION SYSTEM RELIABILITY ANALYSIS
time.” In our analysis, the prescribed time is 6 cycles, which
When applying fault trees to analyze protection system means that this fault tree considers only high-speed protection.
reliability, the power system performance requirements We assume that the power system requires high-speed fault
determine the top event of the fault tree. If, for example, the clearing to preserve transient stability. We should change the
power system requires high-speed fault clearing to preserve prescribed time to breaker failure time if the power system
transient stability, the top event should only consider high- remains stable after a breaker failure protection operation.
speed protection. However, if the power system remains stable The left side of the fault tree in Fig. 26 expresses the
after a breaker failure protection operation, the top event likelihood for the 21 scheme not to provide channel-
should also consider breaker failure protection. independent, high-speed tripping for faults in Zone 1 coverage
We created 20 fault trees (10 for dependability analysis and of both line ends [5]. We assume that both Zone 1 elements
10 for security analysis) for the following protection schemes detect around 45 percent of all line faults to accommodate the
(all with fiber-optic channels): effect of fault resistance. The right side of the fault tree
• Combination of POTT scheme with direct Zone 1 represents POTT/21 scheme contribution to a failure to trip
tripping (POTT/21) in one relay and distance (for the remaining 55 percent of all faults). OR Gate 5
protection (21) scheme in another relay [Fig. 2(a)]. indicates that any failure to trip of the 21 scheme or of the
• Combination of 87L scheme with direct Zone 1 POTT/21 scheme causes a protection scheme failure to trip.
tripping (87L/21) in one relay and 21 scheme in We can modify the fault tree as required to consider other
another relay [Fig. 2(a)]. scheme configurations, include other events of interest, or use
• Dual-redundant POTT/21 scheme [Fig. 2(b)]. Each other unavailability values.
relay performs POTT and 21 scheme functions. In Fig. 26 and in all of the other dependability fault trees,
• Dual-redundant POTT scheme [Fig. 2(b)]. Each relay we show the unavailability values multiplied by 106.
performs POTT scheme functions (no direct Zone 1
tripping).
13
1372
Note: Numbers shown are Protection Fails to Clear

unavailabilities multiplied by 106 In-Section Fault
in the Prescribed Time
0 767
21 Protection Fails to Trip POTT Protection Fails to Trip

in Zone 1 Coverage of NOT in Zone 1 Coverage of
Both S and R Both S and R Breaker Breaker DC CT VT
at S Fails at R Fails System Fails Fails
200 200 Fails 6•9 6 • 15
2 • 30 = 54 = 90
2 3 = 60
447 492
21 Protection R1 at S and R 21 Protection R2 at S and R

Fails to Trip Fails to Trip
in Zone 1 Coverage in Zone 1 Coverage
1394
POTT Protection
Fault in Fails to Trip
Zone 1 of Fault NOT in
Both S and R Overlapping
0.45 Multiplier Zone 1
994 0.55 Multiplier
3 4
497 497 647 647
21 Protection R1 21 Protection R1 POTT Protection POTT Protection

at S Fails at R Fails R1 at S Fails R1 at R Fails
Fiber-Optic
Channel
Same as 21 Same as POTT Fails
Protection R1 at S Protection at S 100
1 2
Relay Relay DC Hidden CT VT Fiber-Optic Comm 21

Fails App or Wiring Failures Wiring Wiring Equip DC Protection
137 Setting Errors 10 Errors Errors Fails System R1 at S
Errors 50 50 50 100 Fails Fails
200 50 497
Fig. 26. Dependability fault tree for the combination of POTT/21 and 21 schemes in separate relays [Fig. 2(a)] with a fiber-optic channel.
14
Fig. 27 shows the security fault tree for the combination of 20 percent of external faults fall within the overreaching
POTT/21 and 21 schemes in separate relays [Fig. 2(a)] with a element zone.
fiber-optic channel. The top event is “protection produces an In Fig. 27 and all of the other security fault trees, we show
undesired trip.” This security fault tree includes the same basic the failure rate values multiplied by 106.
events as the dependability fault tree (Fig. 26) but uses the Fig. 28 shows the dependability fault tree for the
security failure rates shown in Table VI of the Appendix. The combination of 87L/21 and 21 schemes in separate relays
left side of the fault tree expresses the likelihood for the [Fig. 2(a)] with a fiber-optic channel. The left side of the fault
21 scheme to cause an undesired trip. The right side of the tree in Fig. 28 expresses the likelihood for the 21 scheme to
fault tree represents the POTT/21 scheme contribution to an cause a failure to trip. The right side of the fault tree
undesired trip, which occurs when the directly tripping Zone 1 represents the POTT/21 scheme contribution to a failure to
undesirably trips or the communications system generates an trip. OR Gate 6 indicates that any failure to trip of the
undesired permissive trip signal and also an overreaching 21 scheme or of the POTT/21 scheme causes a protection
POTT element operates for an external fault. We assume that scheme failure to trip.
Note: Numbers shown are

16708
failure rates multiplied by 106
Protection Produces an
Undesired Trip
7166
4266
POTT/21 Protection R1
21 Protection R2 at S or R at S or R Produces an
Breaker DC VT Produces an Undesired Trip Undesired Trip
CT
Fails Syst Fails Fails
2 • 333 Fails 6 • 157 6 • 278
= 666 2 • 1000 = 942 = 1668
= 2000
4 5
2133 2133 3583 3583
21 Protection R2 at S 21 Protection R2 at R POTT/21 Protection R1 at S POTT/21 Protection R1 at R

Produces an Undesired Trip Produces an Undesired Trip Produces an Undesired Trip Produces an Undesired Trip
Same as 21 Same as POTT/21

Protection R1 at S Protection R1 at S
2 3
2383 1200
Relay Relay CT VT DC Hidden

21 Prot R1 at S
Fails App or Wiring Wiring Wiring Failures Produces an 0.2
333 Set Errors Errors Errors 50 6000
Undesired Trip
1000 250 250 250
(2133 + 250)
POTT Overreaching
Comm Equipment Produces Elements Operate
Undesired Permissive Trip For External Faults
Comm DC Fiber-Optic Fiber-Optic

System Equip Channel
Fails Fails Fails
2000 2000 2000
Fig. 27. Security fault tree for the combination of POTT/21 and 21 schemes in separate relays [Fig. 2(a)] with a fiber-optic channel.
15
1072

0 558
21 Protection Fails to Trip 87L Protection Fails to Trip

Both S and R Both S and R Breaker Breaker DC CT
at S Fails at R Fails System Fails
200 200 Fails 6•9
2 • 30 = 54
2 3 = 60
488 532

1014
87L Protection
1084 0.55 Multiplier
4 5
542 542 457 457
21 Protection R1 21 Protection R1 87L Protection 87L Protection

Fiber-Optic
Channel
Same as 21 Same as 87L Fails
2 3
307
87L Protection R1 at S Non-

Comm Related Failure
Relay Relay DC Hidden CT VT VT Fiber-Optic Comm
Fails App or Wiring Failures Wiring Wiring Fails Equip DC
137 Setting Errors 10 Errors Errors 3 • 15 Fails System
Errors 50 50 50 = 45 100 Fails
200 50 1
Relay Relay DC Hidden CT

Fails App or Wiring Failures Wiring
137 Setting Errors 10 Errors
Errors 50 50
* Because of the simplicity of 87L settings 60*
Fig. 28. Dependability fault tree for the combination of 87L/21 and 21 schemes in separate relays [Fig. 2(a)] with a fiber-optic channel.
16
Fig. 29 shows the security fault tree for the combination of B. Redundant Schemes
87L/21 and 21 schemes in separate relays [Fig. 2(a)] with a In the redundant schemes analyzed in this paper, either
fiber-optic channel. The left side of the fault tree expresses the Main 1 or Main 2 protection schemes may consist of a
likelihood for the 21 scheme to cause an undesired trip. The POTT/21, POTT, 87L/21, or 87L scheme. We first created the
right side of the fault tree represents the 87L/21 scheme fault trees for these schemes (each residing in one relay) and
contribution to an undesired trip, which occurs when the then used them as building blocks for the fault trees of the
directly tripping Zone 1 undesirably trips or the dual- and triple-redundant schemes. As an example, Fig. 30
communications system generates an undesired 87L trip. shows the dependability fault tree for the POTT/21 scheme in
one relay (Relay R1 in this example).

13410
failure rates multiplied by 106
Undesired Trip
3866 4268
87L/21 Protection R1 at S or R
21 Protection R2 at S or R
Produces an Undesired Trip
Produces an Undesired Trip DC Breaker
CT VT
Fails Fails Syst Trips
6•157 6•278 Fails Undesirably
= 942 = 1668 2•1000 2•333
= 2000 = 666
5 6
1933 1933 2134 2134
21 Protection R2 at S 21 Protection R2 at R 87L/21 Protection R1 at S 87L/21 Protection R1 at R

Produces an Undesired Trip Produces an Undesired Trip Produces an Undesired Trip Produces an Undesired Trip
Same as 21 Same as 87L/21

Protection R1 at S Protection R1 at S
3 4
Relay Relay CT VT DC Hidden Undetected

Fails App or Wiring Wiring Wiring Failures Comm
333 Set Errors Errors Errors 50 6000
Failure
1000 250 250 250 Produces
Comm Equipment Produces 87L
Undesired 87L Assertion External Fault Occurs Undesired
Simultaneous to Trip
Comm Failure 200
0.0001 Multiplier
1933
2
87L/21 Protection R1
Produces Non-Comm
Related Undesired Trip at S
1 Comm DC Fiber-Optic Fiber-Optic

System Equip Channel
Fails Fails Fails
2000 2000 2000
Relay Relay CT VT DC Hidden

Fails App or Wiring Wiring Wiring Failures
333 Set Errors Errors Errors 50
1000 250 250 250
Fig. 29. Security fault tree for the combination of 87L/21 and 21 schemes in separate relays [Fig. 2(a)] with a fiber-optic channel.
17
1279

0 879
21 Protection Fails to Trip POTT Protection Fails to Trip

Both S and R Both S and R Breaker Breaker
at S Fails at R Fails
200 200
2 3
539 584

1598
POTT Protection
1198 0.55 Multiplier
3 4
599 599 749 749
21 Protection R1 21 Protection R1 POTT Protection POTT Protection

Fiber-Optic
Channel
Same as 21 Same as POTT Fails
1 2
Relay Relay DC Hidden DC CT VT CT VT Fiber-Optic Comm 21

Fails App or Wiring Failures System Fails Fails Wiring Wiring Equip DC Protection
137 Setting Errors 10 Fails 3•9 3 • 15 Errors Errors Fails System R1 at S
Errors 50 30 = 27 = 45 50 50 100 Fails Fails
200 50 599
Fig. 30. Dependability fault tree for one of the POTT/21 schemes (Main 1 or Main 2) in one relay in a dual-redundant POTT/21 scheme with a fiber-optic
channel.
Fig. 31 shows how redundancy improves the POTT/21 required to represent systems with lower redundancy (for
scheme dependability. The AND gate indicates that the failure example, single dc power systems or circuit breaker trip coils).
of one scheme does not cause a failure to trip. The In the fault tree shown in Fig. 31, we assume that the
multiplication of unavailabilities reduces the output of the circuit breaker has redundant trip coils, so we split the circuit
AND gate to a value close to zero. We represent full breaker into two parts. We represent circuit breaker trip coil
redundancy in Fig. 31: redundant relays, instrument failures or dc circuit fuse operations at the basic level (below
transformers, dc power systems, communications channels, AND Gate 1). Their contribution to a failure to clear the fault
and circuit breaker trip coils. We can modify this fault tree as is practically eliminated by the AND gate. If the trip coils
18
operate correctly, a breaker failure to interrupt current (a stuck reflects the effect of redundancy: any of the two POTT/21
contact mechanism or a failure of the contacts to extinguish schemes may cause an undesired trip. The result is lower
the arc) will cause a failure to clear the fault, no matter the security (a higher failure rate) than that of the combination of
redundancy of the scheme. Hence, we represent breaker POTT/21 and 21 schemes in separate relays (Fig. 27).
failures to interrupt current above the AND gate as an input to Fig. 33 shows how redundancy improves the 87L/21
OR Gate 1. Because the other input to this OR gate has a very scheme dependability. The AND gate reflects the fact that the
low unavailability value (because of redundancy), the breaker failure of one scheme does not cause a failure to trip. The
failures to interrupt current become the dominant factor in the multiplication of unavailabilities reduces the output of the
scheme dependability. This fact emphasizes the importance of AND gate to a value close to zero. We represent full
good circuit breaker maintenance. It also shows the need for redundancy in Fig. 33: redundant relays, instrument
breaker failure protection that will clear faults in more than transformers, dc power systems, communications channels,
6 cycles but that could be fast enough to prevent the power and circuit breaker trip coils. We can modify this fault tree as
system from losing transient stability. required to represent systems with lower redundancy (for
Fig. 32 shows the security fault tree for the dual-redundant example, single dc power systems or circuit breaker trip coils).
POTT/21 scheme with fiber-optic channels. OR Gate 1
162
unavailabilities multiplied by 106
Protection Fails to Clear
In-Section Fault
81 81
Main 1 and Main 2 Main 1 and Main 2

Protection at S Fails Protection at R Fails
to Clear Fault to Clear Fault
Same as Main 1 and Main 2

Protection at S
1
Main 1 and Main 2

Breaker Protection at S Fails
at S Fails to
Interrupt
Current
80
1119 1169
Main 1 Protection Main 2 protection

at S Fails at S Fails
Main 1: POTT/21 With Main 2: POTT/21 With

Fiber-Optic Channel Fiber-Optic Channel
Fig. 31. Dependability fault tree for the dual-redundant POTT/21 scheme with fiber-optic channels.
19
21718
unavailabilities multiplied by 106 Protection Produces an
Undesired Trip
19052
Breaker Breaker DC
at S Trips at R Trips System
Undesirably Undesirably Fails
333 333 2 • 1000
= 2000
9276 9776
Main 1 Protection Main 2 Protection

Produces an Undesired Trip Produces an Undesired Trip
Main 1: POTT/21 With Fiber-Optic Channel Main 2: POTT/21 With Fiber-Optic Channel
Fig. 32. Security fault tree for the dual-redundant POTT/21 scheme with fiber-optic channels.
Note: Numbers shown are 162

Protection Fails to Clear
In-Section Fault
81 81
Main 1 and Main 2 Main 1 and Main 2
Protection at S Fails Protection at R Fails
to Clear Fault to Clear Fault
Same as Main 1 and Main 2

Protection at S
1
Main 1 and Main 2

Protection at S Fail
Breaker
at S Fails to
Interrupt
Current
80 1
912 927

Main 1: 87L/21 With Fiber-Optic Channel Main 2: 87L/21 With Fiber-Optic Channel
Fig. 33. Dependability fault tree for the dual-redundant 87L/21 scheme with fiber-optic channels.
20
Fig. 34 shows the security fault tree for the dual-redundant reliability is practically the same. Tripping occurs when at
87L/21 scheme with fiber-optic channels. OR Gate 1 reflects least two of the schemes operate. The effect of the voting logic
the effect of redundancy: any of the two 87L/21 schemes may is that the output of AND Gate 2 is practically zero (very high
cause an undesired trip. The result is lower security (a higher dependability). Hence, breaker failures to interrupt current
failure rate) than that of the 87L/21 and 21 schemes in determine the scheme dependability. In this analysis, we
separate relays (Fig. 29). assume that the three schemes have the same RF coverage. If
Fig. 35 shows the dependability fault tree for a triple- the schemes had different RF coverages (because of different
redundant two-out-of-three voting 87L/21 scheme with fiber- settings, principles of operation, or manufacturers) and two of
optic channels. The scheme has three independent 87L/21 the schemes did not detect a high-resistance in-section fault,
schemes (three relays, three communications channels, three the two-out-of-three voting scheme would fail to clear the
sets of instrument transformers, three dc power systems, and fault. Such a combination of schemes may consist, for
circuit breakers with three trip coils). Actually, circuit example, of two 87L schemes using only 87LP elements with
breakers have only two trip coils. However, if the voting a third 87L scheme that includes 87LP and 87LQ elements.
scheme sends the tripping signal to both trip coils, the fault For this reason, we recommend that voting schemes use relays
tree is slightly different from that of Fig. 35 but the scheme with the same RF coverage.
Note: Numbers shown are 17722
Undesired Trip
15056
Breaker Breaker DC
at S Trips at R Trips System
Undesirably Undesirably Fails
333 333 2 • 1000
= 2000
7278 7778

Main 1: 87L/21 With Fiber-Optic Channel Main 2: 87L/21 With Fiber-Optic Channel
Fig. 34. Security fault tree for the dual-redundant 87L/21 scheme with fiber-optic channels.
21
160
unavailabilities multiplied by 106 Protection Fails to Clear
In-Section Fault in the
Prescribed Time
80 80
Protection at S Protection at R
Fails to Clear Fault Fails to Clear Fault
Same as
Protection at S
1
Protection at S Fails
Breaker at S
Fails to
Interrupt
Current
80 2
1 1 1
Main 1 and Main 2 Main 1 and Main 3 Main 2 and Main 3

Protection at S Fails Protection at S Fails Protection at S Fails
912 927

Fig. 35. Dependability fault tree for the triple-redundant two-out-of-three voting 87L/21 scheme with fiber-optic channels.
22
3008
failure rates multiplied by 106 Protection Produces an
Undesired Trip
1504 1504
Protection at S Protection at R
Same as
Protection at S
1
57 57 57
Main 1 and Main 2 Main 1 and Main 3 Main 2 and Main 3
Protection at S Produces Protection at S Produces Protection at S Produces
an Undesired Trip an Undesired Trip an Undesired Trip
DC Breaker
System Trips
Fails Undesirably
1000 333 1
7278 7778
Main 1 Protection at S Main 2 Protection at S

Fig. 36. Security fault tree for the triple-redundant two-out-of-three voting 87L/21 scheme with fiber-optic channels.
Fig. 36 shows the security fault tree for the triple-redundant

two-out-of-three voting 87L/21 scheme with fiber-optic
channels. AND Gate 1 reflects the fact that two schemes need
to misoperate to cause an undesired trip. The result is very
high security (a low failure rate).
Table VII of the Appendix summarizes the results obtained
from the fault tree analysis. Fig. 37 and Fig. 38 present these
results in a graphical form. Lower values in these figures
mean higher reliability.
Fig. 38. Fault tree analysis results: Line protection security comparison.
VII. PROTECTION SYSTEM SELECTIVITY ANALYSIS

As mentioned before, selectivity is the ability of a
protection system to eliminate a fault in the shortest time
possible with the least disconnection of system components.
By striving for high-speed tripping on all line terminals while
ensuring high levels of security and dependability, we
essentially achieve selectivity.
Fig. 37. Fault tree analysis results: Line protection dependability
comparison.
23
VIII. PRACTICAL CONSIDERATIONS B. Mutual Coupling [21]

In this section we discuss the line protection challenges Magnetic mutual coupling affects ground directional
presented by series-compensated, mutually coupled, overcurrent elements responding to zero-sequence quantities,
multiterminal, and short lines. We also discuss distance which compromises directional comparison scheme security.
element problems when applied to extra-high-voltage (EHV) Mutual coupling may cause zero-sequence polarizing quantity
lines that require single-pole tripping and reclosing. In inversion when this coupling is strong enough to dominate
addition, we describe the advantages of 87L protection in over the electrical connection between lines. An extreme case
dealing with these complex transmission line protection is when the zero-sequence network of the protected line is
applications. electrically isolated from the zero-sequence network of the
faulted line. Zero-sequence polarized directional elements can
A. Series Compensation [18] [19]
misoperate for reverse faults under certain system
Series-compensated lines present unique challenges for configurations and circuit breaker switching conditions. A
directional, distance, and 87L elements because the transient solution to this problem is to use 67Q elements.
response of the series capacitor is not readily predictable. Magnetic mutual coupling affects ground distance elements
Series compensation introduces errors in the impedance that and compromises distance and directional comparison scheme
distance elements estimate. The series capacitor modifies the security and dependability. Ground distance elements
line impedance that the relay measures. Furthermore, overreach (a concern for Zone 1) when the zero-sequence
subharmonic frequency oscillations cause the impedance currents in the protected line and the coupled line flow in
estimation to oscillate. The basic problem is that the opposite directions and underreach (a concern for Zone 2)
impedance estimation depends on the state of the capacitor when these currents flow in the same direction. Solutions to
protection. this problem include applying reach or zero-sequence
A voltage inversion is a change of approximately compensation settings that consider the mutual coupling
180 degrees in the voltage phase angle. For elements effect.
responding to phase quantities, voltage inversion can occur for 87L protection is not affected by mutual coupling and is an
a fault near a series capacitor if the impedance from the relay excellent solution for mutually coupled lines.
to the fault is capacitive rather than inductive. Voltage Because of mutual coupling, operating a double-circuit
inversion can cause directional and distance elements to transmission line as a single circuit with jumpers placed across
operate incorrectly. Close-in, three-phase, bolted faults present similar phases along the line causes phase and ground distance
an additional problem to directional and distance elements element underreaching [21]. Applying 87L schemes solves
responding to phase voltages because these elements lose their this problem.
polarizing voltage. In series-compensated lines, the voltage at
the fault side of the series capacitor collapses and voltage C. Single-Pole Tripping [22] [23]
reverses on the other side. In particular, external faults may Single-pole tripping and reclosing protection schemes are
appear to be internal, and internal faults may appear to be designed to trip only the faulted phase on single-phase-to-
external. Memory polarization and offset characteristics in ground faults and all three phases on multiphase faults. When
directional elements can make distance and directional a single-phase-to-ground fault occurs on a transmission line,
elements secure for voltage inversions. the faulted phase is tripped and automatically reclosed after a
A current inversion occurs on a series-compensated line suitable dead time, which should be long enough for the
when, for an internal fault, the equivalent system at one side secondary arc caused by the coupling with the unfaulted
of the fault is capacitive and the equivalent system at the other phases to extinguish. If the fault is cleared and of transient
side of the fault is inductive. The current flows out of the line nature, the scheme resets. If the fault is still present when the
at one terminal, which is referred to as current outfeed. For pole is reclosed, all three poles are tripped and no further
high-resistance faults, the low fault current prevents capacitor reclosing takes place. When a multiphase fault occurs on a
bypassing and creates conditions for a current inversion. transmission line, all three poles are tripped and typically no
Current inversion can also occur in negative- or zero-sequence reclosing takes place.
networks. Current inversion affects directional, distance, A single-pole tripping scheme must distinguish between
phase comparison, and differential elements responding to multiphase faults and single-phase-to-ground faults. For
phase or sequence-component quantities. Current inversions multiphase faults, the scheme trips three circuit breaker poles.
are unlikely, but they may occur under certain conditions. For single-phase-to-ground faults, the scheme produces a
87L protection is an excellent choice for series- single-pole tripping output associated with the faulted phase.
compensated lines. 87L elements are immune to voltage For this reason, the scheme requires a faulted phase
inversions and more tolerant of current inversions than are identification logic.
directional or distance elements. In particular, the Alpha Plane The open-phase condition following a single-pole trip on a
87L elements described in [20] are very tolerant of current transmission line creates unbalances that can affect relays. The
inversion and subharmonic frequency transients. protection elements must be designed to be immune to the
unbalance effects or must be desensitized or blocked during
the single-pole open time. In addition, single-pole tripping
24
schemes must detect faults that occur or evolve during the from the other terminals. The infeed increases the measured
single-pole open time. For this reason, the location of the impedance value and causes distance elements to underreach.
voltage transformers (VTs) is very important in schemes based In some three-terminal applications, there may be an
on distance or directional overcurrent elements. VTs located outfeed current at one of the terminals during an internal line
on the line side of the circuit breaker present challenges to fault rather than an infeed current [3]. In this case, the
distance and directional elements. measured impedance at the other line terminals may be
Polarizing quantity corruption can occur during single-pole smaller than the actual impedance to the fault (distance
open conditions in applications with line-side VTs if one of elements tend to overreach). An additional problem is that a
the input voltages to the memory polarizing algorithm is forward-looking distance element at the terminal with outfeed
corrupted. Incorrect memory polarization may cause distance current will not detect this internal line fault. In fact, if there is
element misoperation. The voltage magnitude and angle on a blocking element at the terminal with outfeed current, it may
the line side of the circuit breaker during single-pole open declare the internal fault as an external fault and prevent
conditions depend on whether a secondary arc still exists on tripping in a DCB scheme. If the Zone 1 elements at one of the
the open phase, whether line shunt reactors are present, and other terminals respond to the fault, they will initiate circuit
whether power flows in the two healthy phases. breaker tripping and remove the outfeed current. The
Shunt reactors located on the line side of the circuit remaining line terminals may then trip via the directional
breakers compensate the line charging currents and reduce comparison scheme; however, tripping may be delayed by the
overvoltages in long transmission lines. When the circuit current reversal logic.
breakers open at both line ends, the remaining circuit is In many three-terminal line applications, the third terminal
basically an RLC circuit with stored energy in the reactor may be a transformer-terminated load tap. In such
inductance and the line capacitance. The shunt reactors applications, there may be no positive- or negative-sequence
interact with the line capacitance and maintain ringing line current source at the tap. However, if the line side of the
voltages for several cycles. With line-side VTs, these ringing transformer has a grounded-wye connection, there will be a
voltages corrupt the distance protection polarization and significant zero-sequence current contribution. Therefore, in
frequency estimation. this application, the infeed effect does not affect the phase
An evolving fault may start as a single-phase-to-ground distance elements but it does affect the ground distance
fault and then involve additional phases while the initial fault elements.
is being cleared or during the reclosing dead time of the 87L schemes for multiterminal lines handle three-terminal
original faulted phase. Single-pole tripping schemes should line protection challenges naturally. In addition, 87L schemes
detect and clear evolving faults. can handle some level of outfeed current, depending on the
In summary, single-pole tripping schemes must identify the relay operating characteristic and the applied relay settings.
faulted phases, avoid misoperation on unbalances created by
E. Short Lines
the open phase condition, and detect faults that occur or
evolve during the single-pole open time. These requirements Transmission lines are sometimes classified as short,
increase the complexity of the logic of single-pole tripping medium, and long. The IEEE Power and Engineering Society
schemes with distance and directional overcurrent elements. (PES) Power System Relaying Committee (PSRC) established
87L schemes perform well during the single-pole open time the following criteria for line length in terms of the SIR [25]:
and detect evolving faults naturally, because they compare the Short line: SIR > 4.0.
line terminal currents on a per-phase basis. Medium line: 0.5 < SIR < 4.0.
Long line: SIR < 0.5.
D. Multiterminal and Tapped Lines Short line protection problems related to the high SIR value
Multiterminal lines have three or more terminals, each with include low voltage at the relay location during faults and
substantial generation. Distance protection application to CCVT transients. Detecting high-resistance faults is another
three-terminal lines is more complex than its application to short line protection problem.
two-terminal lines because of the large variety of possible tap Microprocessor-based relays do not require a considerable
locations, line impedances, source impedances, and system amount of energy to operate. However, when the relay input
operation conditions [24]. voltage is very low, distance element operation is not well-
We need to determine the impedance measured by distance defined and/or the operating speed is unsatisfactory. For
elements for various fault and system conditions when proper distance element performance during faults, the voltage
calculating their settings. The measured impedance is not measured by the relay must be above the design voltage
always the actual impedance of the line section from the relay threshold. The SIR value determines the positive-sequence
location to the fault point. The voltage measured by the relay voltage at the relay location for a three-phase fault at the relay
is the voltage drop between its location and the fault, which is reach point. As the SIR increases, the voltage at the relay
affected by the infeed effect from the sources connected at the location decreases, which limits distance element sensitivity in
taps between the relay and the fault location. Thus, the short lines.
measured impedance depends on the current contributions
25
Coupling capacitor voltage transformer (CCVT) transients IX. CONCLUSION

may impair the voltage measured by the relay during faults This paper compares several line protection schemes,
and cause distance elements to overreach. A Zone 1 including distance schemes, directional comparison schemes
overreaching condition due to CCVT transients is generally using distance and directional elements, and 87L schemes in
more likely in short lines (because of the higher SIR) than in terms of speed, sensitivity, reliability, and selectivity. From
long lines [26]. For high SIR values, the CCVT transient lasts this comparison, we conclude the following:
longer and is more severe because the voltage at the CCVT • The operating speed of phasor-based protection
location is low for remote faults. Also, the CCVT transient elements depends mainly on the window length of the
can distort the voltage waveform because most of the voltage digital band-pass filter. Modern phasor-based line
signal consists of the CCVT transient output [3]. For low SIR protection elements with half-cycle or one-cycle data
values, the voltage at the CCVT location is high for remote windows typically provide PSTT values of one cycle
faults and the CCVT transient has little effect on the voltage or less without impairing security.
waveform. The traditional solutions (i.e., reducing Zone 1 • In line pilot protection schemes, the communications
reach or delaying its operation) affect high-speed fault system delays operation. A relay-to-relay
clearing. communications channel over a direct optical fiber
Modern microprocessor-based relays include logic to introduces the smallest delay (0.8 ms/100 miles). In
prevent the Zone 1 element from operating because of CCVT digital multiplexed fiber-optic networks, the
transients [3]. If CCVT blocking logic is enabled and the relay multiplexer adds approximately 0.5 ms and each
detects a high SIR value when a Zone 1 element picks up, the repeater (if used) adds 24 µs. A PLC channel adds a
logic delays tripping for as long as 1.5 cycles to allow the delay between 2 and 8 ms. Protection-class digital
CCVT transient to stabilize. For each fault, the relay estimates radios introduce delays between 4.8 and 7.4 ms.
the SIR value as the ratio of the positive-sequence source
• Computer simulation studies of three different
impedance to the Zone 1 reach setting. The logic does not
example power systems show that modern protection
require settings. If the distance calculation stabilizes before
schemes using high-speed elements and fast
1.5 cycles, the logic unblocks tripping. Therefore, Zone 1
communications channels produce low (always under
elements operate without significant delay when the CCVT
1.2 cycles) and consistent PSTT values for various
has good transient response. A better solution is 87L
fault types and locations, line lengths, and source
protection, which does not use voltage information.
strengths. Many actual system events have validated
Another consideration for short line protection is RF
the results of these simulations. Time-domain
coverage. For ground faults, the total fault resistance can be
protection can further improve speed.
very high, such as in the case of a fault involving a tree, a fault
• Plotting the maximum value of RF detected by a line
to ground through a fire, or a fault on a tower with very high
protection scheme as a function of the fault distance
footing resistance. The infeed effect from the remote line
allows us to simultaneously analyze the scheme
terminal current on the resistive fault path magnifies the
sensitivity and speed. The RF coverage regions
measured fault impedance and shifts its phase angle. The
resulting from computer simulation studies of three
apparent fault impedance may be much higher than the line
different example power systems show the higher
impedance. A ground distance element may not detect these
sensitivity and speed of 87L schemes using 87LG and
high-resistance faults. A better solution is ground directional
87LQ elements as compared with POTT schemes
overcurrent protection (67N) or 87L protection with 87LG or
using either 67N or 21N (or 21X) elements. These
87LQ elements [26].
studies also show that 67N elements provide higher
Even for phase faults, the measured fault impedance may
sensitivity and speed than 21N (or 21X) elements.
be high, as compared with the line impedance, and may limit
• From the dependability and security fault trees for
phase distance element sensitivity.
different line pilot protection schemes with fiber-optic
F. Advantages of 87L Protection [3] channels, we conclude the following:
87L protection schemes do not require voltage information, − The combination of 87L/21 and 21 schemes in
thereby avoiding problems for close-in faults, blown potential separate relays provides higher dependability and
fuses, ferroresonance in VTs, transients in CCVTs, and security than the combination of POTT/21 and 21
voltage inversion. However, 87L elements may require schemes in separate relays.
voltage information to calculate the line charging current in − The dual-redundant schemes with direct Zone 1
applications for long lines or cables. 87L protection schemes tripping (dual POTT/21, POTT/21 and 87L/21, and
are almost immune to unbalances, current reversals on parallel dual 87L/21 schemes) have much higher
lines, power swings, and magnetic mutual coupling. In dependability and slightly lower security than the
addition, they perform well for evolving, intercircuit, and corresponding combinations of schemes in separate
cross-country faults; tolerate high line loading; and may relays.
handle outfeed conditions, depending on their operating − Removing Zone 1 tripping from the dual-redundant
characteristic. schemes improves their security without impairing
26
their dependability. However, direct Zone 1 speed, sensitivity, dependability, security, and
tripping is faster than communications-based selectivity.
schemes (there is no channel delay) and enhances
power system stability by providing faster clearing X. APPENDIX
of close-in faults. Table V summarizes the results of the reliability study
− The triple-redundant voting 87L/21 scheme is using fault trees reported in [1].
significantly more secure than the dual-redundant Table VI summarizes the reliability indices used in the
87L/21 and 87L schemes. Removing Zone 1 fault tree analysis. We used the reliability indices reported in
tripping (triple-redundant voting 87L scheme) [1] (which provides a detailed justification of these indices),
further improves security. with one exception. We modified the value of unavailability
• A protection scheme that provides high-speed tripping caused by relay application and setting errors based on a
on all line terminals and high levels of security and recent North American Electric Reliability Corporation
dependability inherently provides selectivity. (NERC) report that provides protection system misoperation
• 87L protection is the best solution for lines. 87L data for a fifteen-month period [27]. NERC concludes that
protection does not require voltage information and is incorrect settings, logic, and design errors caused 28 percent
almost immune to unbalances, current reversals on of the misoperations, while relay failures caused 20 percent of
parallel lines, power swings, and magnetic mutual the misoperations. Using this information, we assumed
coupling. 87L protection performs well for evolving, U = 200 • 10–6 for relay application or settings errors (instead
intercircuit, and cross-country faults; tolerates high of the U = 1,000 • 10–6 value used in [1]).
line loading; and may handle outfeed conditions, Table VII summarizes the results of the reliability study
depending on its operating characteristic. using fault trees described in this paper.
• Triple-redundant voting 87L and dual-redundant 87L
protection schemes allow maximizing line protection
TABLE V
LINE PROTECTION RELIABILITY COMPARISON [1]
Dependability Security
(Unavailability • 106) (Failure Rate • 106)
Protection Scheme Normal Comprehensive Normal Comprehensive
Commissioning Commissioning Commissioning Commissioning
Testing Testing Testing Testing
Basic POTT (microwave) 2,562 1,339 (1.9 times) 23,318 12,938 (1.8 times)
Basic POTT (optical fiber) 2,452 1,229 (2.0 times) 22,784 12,364 (1.8 times)
Basic DCB (PLC) 2,122 943 (2.3 times) 48,704 33,180 (1.5 times)
Dual-redundant POTT 168 162 (1.04 times) 27,052 16,072 (1.7 times)
Dual-redundant POTT with relays from different manufacturers 174 162 (1.07 times) 29,552 16,572 (1.8 times)
Dual-redundant POTT with common-mode failures 1,178 268 (4.4 times) 28,102 16,202 (1.7 times)
Fully redundant voting POTT 160 160 (1.0 times) 916 750 (1.2 times)
Voting POTT: Two schemes share dc power system 220 172 (1.3 times) 2,892 1,146 (2.5 times)
Voting POTT: Two schemes also share communications channel 1,120 992 (1.1 times) 6,592 4,224 (1.6 times)
Voting POTT: Two schemes also share instrument transformers 1,464 1,136 (1.3 times) 10,182 6,826 (1.5 times)
Fully redundant voting POTT with common-mode failures 1,170 266 (4.4 times) 1,966 880 (2.2 times)
Note: The numbers in parentheses represent the effect of comprehensive commissioning testing. These numbers are the ratios of the unavailabilities or failure
rates with normal testing to the unavailabilities or failure rates with comprehensive testing.
27
TABLE VI
RELIABILITY INDICES USED IN FAULT TREES [1]
Event
Unavailability • 106 MTBF (Years) Failure Rate • 106
Relay fails 137 3,000 333
Relay application or settings errors 200 1,000 1,000
Circuit breaker fails 200 3,000 333
Circuit breaker fails to interrupt current 80 – –
DC power system fails 30 1,000 1,000
Current transformer (CT) fails 9 6,370 157
VT fails 15 3,600 278
Fiber-optic equipment fails 100 500 2,000
Fiber-optic channel fails 100 500 2,000
Communications dc power system fails 50 500 2,000
DC system wiring errors 50 4,000 250
CT or VT wiring errors 50 4,000 250
Hidden failures 10 20,000 50
Note: The U = 200 • 10–6 value for relay application or settings errors shown in this table is different from the value reported in [1].
TABLE VII
LINE PROTECTION RELIABILITY COMPARISON
Protection Scheme
(Unavailability • 106) (Failure Rate • 106)
Combination of POTT/21 and 21
1,372 16,708
in separate relays
Combination of 87L/21 and 21
1,072 13,410
in separate relays
Dual-redundant POTT/21 162 21,718
Dual-redundant POTT 168 15,942
Dual-redundant POTT/21 and 87L/21 162 19,220
Dual-redundant POTT and 87L 166 12,864
Dual-redundant 87L/21 162 17,722
Dual-redundant 87L 164 9,786
Triple-redundant voting 87L/21 160 3,008
Triple-redundant voting 87L 160 2,744
XI. REFERENCES [4] P. M. Anderson, Power System Protection. New York: McGraw-Hill,
1999.
[1] E. O. Schweitzer, III, D. Whitehead, H. J. Altuve Ferrer, D. A.
Tziouvaras, D. A. Costello, and D. Sánchez Escobedo, “Line Protection: [5] E. O. Schweitzer, III, B. Fleming, T. J. Lee, and P. M. Anderson,
Redundancy, Reliability, and Affordability,” proceedings of the 37th “Reliability Analysis of Transmission Protection Using Fault Tree
Annual Western Protective Relay Conference, Spokane, WA, October Methods,” proceedings of the 24th Annual Western Protective Relay
2010. Conference, Spokane, WA, October 1997.
[2] R. Sandoval, C. A. Ventura Santana, H. J. Altuve Ferrer, R. A. [6] E. O. Schweitzer, III, and D. Hou, “Filtering for Protective Relays,”
Schwartz, D. A. Costello, D. A. Tziouvaras, and D. Sánchez Escobedo, proceedings of the 19th Annual Western Protective Relay Conference,
“Using Fault Tree Analysis to Evaluate Protection Scheme Spokane, WA, October 1992.
Redundancy,” proceedings of the 37th Annual Western Protective Relay [7] D. Hou, A. Guzmán, and J. Roberts, “Innovative Solutions Improve
Conference, Spokane, WA, October 2010. Transmission Line Protection,” proceedings of the 24th Annual Western
[3] H. J. Altuve Ferrer and E. O. Schweitzer, III (eds.), Modern Solutions Protective Relay Conference, Spokane, WA, October 1997.
for Protection, Control, and Monitoring of Electric Power Systems. [8] A. Guzmán, J. Mooney, G. Benmouyal, N. Fischer, and B. Kasztenny,
Schweitzer Engineering Laboratories, Inc., Pullman, WA, 2010. “Transmission Line Protection System for Increasing Power System
Requirements,” proceedings of the 55th Annual Conference for
Protective Relay Engineers, College Station, TX, April 2002.
28
[9] E. O. Schweitzer, III, B. Kasztenny, A. Guzmán, V. Skendzic, and M. XII. BIOGRAPHIES

V. Mynam, “Speed of Line Protection – Can We Break Free of Phasor
Limitations?” proceedings of the 41st Annual Western Protective Relay Héctor J. Altuve received his B.S.E.E. degree in 1969 from the Central
University of Las Villas in Santa Clara, Cuba, and his Ph.D. degree in 1981
Conference, Spokane, WA, October 2014.
from Kiev Polytechnic Institute in Kiev, Ukraine. From 1969 until 1993, Dr.
[10] Ametek Power Instruments, “TCF-10B Specifications.” Available: Altuve served on the faculty of the Electrical Engineering School at the
www.ametekpower.com/download/TCF-10B-Data-Sheet.pdf Central University of Las Villas. From 1993 to 2000, he served as professor
[11] K. Behrendt and K. Fodero, “Implementing Mirrored Bits Technology of the Graduate Doctoral Program in the Mechanical and Electrical
Over Various Communications Media,” SEL Application Guide 2001- Engineering School at the Autonomous University of Nuevo León in
12. Available: https://www.selinc.com/literature/ApplicationGuides. Monterrey, Mexico. In 1999 through 2000, he was the Schweitzer Visiting
[12] S. Achanta, R. Bradetich, and K. Fodero, “Speed and Security Professor in the Department of Electrical and Computer Engineering at
Washington State University. Dr. Altuve joined Schweitzer Engineering
Considerations for Protection Channels,” proceedings of the 42nd
Laboratories, Inc. (SEL) in January 2001, where he is currently a
Annual Western Protective Relay Conference, Spokane, WA, October
distinguished engineer and dean of SEL University. He has authored and
2015. coauthored more than 100 technical papers and several books and holds four
[13] T. Jiao and C. F. Henville, “Evolving Transmission Line Faults While patents. His main research interests are in power system protection, control,
Single Phase Open,” proceedings of the 41st Annual Western Protective and monitoring. Dr. Altuve is an IEEE fellow.
Relay Conference, Spokane, WA, October 2014.
[14] J. Roberts, E. O. Schweitzer, III, R. Arora, and E. Poggi, “Limits to the Karl Zimmerman is the technical support director at Schweitzer Engineering
Sensitivity of Ground Directional and Distance Protection,” proceedings Laboratories, Inc. in Fairview Heights, Illinois. His work includes providing
of the 22nd Annual Western Protective Relay Conference, Spokane, application and product support and technical training for protective relay
WA, October 1995. users. He is a senior member of IEEE, a member of the IEEE Power System
[15] J. Mooney and J. Peer, “Application Guidelines for Ground Fault Relaying Committee, and vice-chairman of the line protection subcommittee.
Protection”, proceedings of the 52nd Annual Georgia Tech Protective Karl received his B.S.E.E. degree at the University of Illinois at Urbana-
Relaying Conference, Atlanta, GA, May 1998. Champaign and has over 20 years of experience in the area of system
protection. He has authored over 25 papers and application guides on
[16] E. O. Schweitzer, III, and J. J. Kumm, “Statistical Comparison and protective relaying and was honored to receive the 2008 Walter A. Elmore
Evaluation of Pilot Protection Schemes,” proceedings of the 23rd Best Paper Award from the Georgia Institute of Technology Protective
Annual Western Protective Relay Conference, Spokane, WA, October Relaying Conference.
1996.
[17] P. K. Maezono et al., “Very High-Resistance Fault on a 525 kV Demetrios A. Tziouvaras is a professor at SEL University. He joined
Transmission Line – Case Study,” proceedings of the 35th Annual Schweitzer Engineering Laboratories, Inc. (SEL) in 1998 and has over
Western Protective Relay Conference, Spokane, WA, October 2008. 35 years of power system protection experience. Demetrios researched
[18] D. A. Tziouvaras and A. Apostolov, “Experience With Directional advanced relaying algorithms for 14 years at SEL prior to joining the SEL
Comparison Protection for Series-Compensated Transmission Lines,” University. He previously worked at Pacific Gas and Electric Company for
proceedings of the 25th Annual Western Protective Relay Conference, 18 years, where he held various protection engineering positions, including
Spokane, WA, October 1998. principal protection engineer. He holds five patents in the area of power
system protection and has authored and coauthored more than 60 technical
[19] H. J. Altuve, J. B. Mooney, and G. E. Alexander, “Advances in Series- papers. He is an IEEE senior member and a member of the Power System
Compensated Line Protection,” proceedings of the 62nd Annual Relaying Committee (PSRC). He is a member of CIGRE and an executive
Conference for Protective Relay Engineers, College Station, TX, March member of the U.S. National Committee of CIGRE. He received his B.S.E.E.
2009. from the University of New Mexico and M.S.E.E. from Santa Clara
[20] J. Roberts, D. Tziouvaras, G. Benmouyal, and H. J. Altuve, “The Effect University. His is active in several IEEE PSRC and CIGRE working groups.
of Multiprinciple Line Protection on Dependability and Security,” His main research interests are in power system protection, control, and
proceedings of the 55th Annual Protective Relaying Conference, monitoring.
Atlanta, GA, May 2001.
[21] D. A. Tziouvaras, H. J. Altuve, and F. Calero, “Protecting Mutually
Coupled Transmission Lines: Challenges and Solutions,” proceedings of
the 40th Annual Western Protective Relay Conference, Spokane, WA,
October 2013.
[22] E. Godoy, A. Celaya, H. J. Altuve, N. Fischer, and A. Guzmán,
“Tutorial on Single-Pole Tripping and Reclosing,” proceedings of the
39th Annual Western Protective Relay Conference, Spokane, WA,
October 2012.
[23] D. A. Tziouvaras (convener), “Modern Distance Protection Functions
and Applications,” CIGRE Technical Brochure 359, October 2008.
[24] G. E. Alexander, “Applying the SEL-311C Relay on Three Terminal
Lines,” SEL Application Guide 2000-12. Available: https://www.selinc.
com/literature/ApplicationGuides.
[25] ANSI/IEEE Standard C37.113-1999, IEEE Guide for Protective Relay
Applications to Transmission Lines.
[26] G. E. Alexander, J. G. Andrichak, and W. Z. Tyska, “Relaying Short
Lines,” proceedings of the 44th Annual Conference for Protective Relay
Engineers, College Station, TX, April 1991.
[27] North American Electric Reliability Corporation, “Misoperations
Report,” prepared by the Protection System Misoperation Task Force,
April 2013. Available: http://www.nerc.com/docs/pc/psmtf/PSMTF_
Report.pdf
© 2015 by Schweitzer Engineering Laboratories, Inc.
All rights reserved.
20150910 • TP6711-01

Paper For Research Topic MS Stduents

Uploaded by

Copyright:

Available Formats

Paper For Research Topic MS Stduents

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Paper For Research Topic MS Stduents

Uploaded by

Copyright:

Available Formats

Maximizing Line Protection Reliability,

Speed, and Security

Héctor J. Altuve, Karl Zimmerman, and Demetrios Tziouvaras

Maximizing Line Protection Reliability,

Bus S Bus R B. Example Protection Systems

Fiber-Optic Digital Fiber- Fiber-Optic

125 Vdc 48 Vdc 48 Vdc 125 Vdc

125 Vdc 2 48 Vdc 2 48 Vdc 2 125 Vdc 2

Fiber-Optic Digital Fiber- Fiber-Optic

Fiber-Optic Digital Fiber- Fiber-Optic

125 Vdc 1 48 Vdc 1 48 Vdc 1 125 Vdc 1

protection algorithms: directional overcurrent, distance, and

cycle, and so on) directly impacts relay operating speed and

can combine long- and short-window filters to achieve faster 0.50

High-Speed Elements B. Pilot Protection Schemes

1.0 • Direct underreaching transfer trip (DUTT).

Communications Communications latency. Modern communications equipment uses high-speed,

(a) Relay Processing Delay (ms)

Communications Communications 38.4 4.2 8.3

TABLE II Delay (ms)

0 0.25 0.50 0.75 1.0 0 0.25 0.50 0.75 1.0

0 0.25 0.50 0.75 1.0

0 0.25 0.50 0.75 1.0

1:TRIP 21N 21N

− Resistance reach setting of 50 ohms secondary.

Fig. 21 shows the RF coverage regions of a POTT scheme 125

regions (corresponding to Region A in Fig. 21) for the 67Q

C. Actual Fault Case • Dual-redundant POTT/21 and 87L/21 scheme

Note: Numbers shown are Protection Fails to Clear

21 Protection Fails to Trip POTT Protection Fails to Trip

21 Protection R1 at S and R 21 Protection R2 at S and R

497 497 647 647

21 Protection R1 21 Protection R1 POTT Protection POTT Protection

Relay Relay DC Hidden CT VT Fiber-Optic Comm 21

Note: Numbers shown are

2133 2133 3583 3583

21 Protection R2 at S 21 Protection R2 at R POTT/21 Protection R1 at S POTT/21 Protection R1 at R

Same as 21 Same as POTT/21

Relay Relay CT VT DC Hidden

Comm DC Fiber-Optic Fiber-Optic

Note: Numbers shown are Protection Fails to Clear

21 Protection Fails to Trip 87L Protection Fails to Trip

21 Protection R1 at S and R 21 Protection R2 at S and R

542 542 457 457

21 Protection R1 21 Protection R1 87L Protection 87L Protection

87L Protection R1 at S Non-

Relay Relay DC Hidden CT

Note: Numbers shown are

1933 1933 2134 2134

21 Protection R2 at S 21 Protection R2 at R 87L/21 Protection R1 at S 87L/21 Protection R1 at R

Same as 21 Same as 87L/21

Relay Relay CT VT DC Hidden Undetected

1 Comm DC Fiber-Optic Fiber-Optic

Relay Relay CT VT DC Hidden

Note: Numbers shown are Protection Fails to Clear

21 Protection Fails to Trip POTT Protection Fails to Trip

21 Protection R1 at S and R 21 Protection R2 at S and R

599 599 749 749