Protection and Testing Considerations for

IEC 61850 Sampled Values-Based Distance

and Line Current Differential Schemes

Steven Chase, Erin Jessup, Mauricio Silveira, Jiawei Dong, and Qiaoyin Yang
Schweitzer Engineering Laboratories, Inc.

Presented at the
72nd Annual Conference for Protective Relay Engineers
College Station, Texas
March 25–28, 2019
 1

Protection and Testing Considerations for

IEC 61850 Sampled Values-Based Distance
and Line Current Differential Schemes
Steven Chase, Erin Jessup, Mauricio Silveira, Jiawei Dong, and Qiaoyin Yang,
Schweitzer Engineering Laboratories, Inc.

Abstract—The implementation of Sampled Values-based (SV- these factors into account, and sufficient testing is necessary to
based) protection and control systems requires that new determine which network issues exist and how severe they are.
equipment be installed in the substation that is not required with This paper shows computer simulations of communications
traditional protection systems. An SV-based system includes
merging units (MUs) that convert analog signals to SV, Ethernet conditions and explains how these conditions impact the SV
network switches, a high-accuracy time source, fiber-optic cables, messages received by an SV relay. We also discuss the impact
and SV relays, all connected to a communications network. When of these conditions on line protection and the considerations
designing SV-based substations, engineers must learn about that protection engineers should factor in when selecting
communications conditions in the system and their effects on settings for SV relays.
relaying applications. Protection engineers often have concerns To investigate the effects of communications conditions on
about whether they should test for delays introduced by a
communications network in their protection system and about line distance protection, we propose a closed-loop test model to
what other effects they need to consider. Ultimately, they need to perform benchmark testing of SV-based schemes. In this test, a
prove that SV-based protection schemes are comparable to power system simulation tool models a long transmission line.
traditional protection systems. The current transformer (CT) and voltage transformer (VT)
This paper discusses communications conditions, such as signals generated by this tool are fed to a merging unit (MU),
bandwidth limitations, latency, and packet loss, and analyzes them published as SV messages, and subscribed to by an SV distance
with respect to SV-based protection. We examine the impacts of
relay. A traditional distance relay that receives CT and VT
SV data loss on line percentage differential, Alpha Plane
differential, and line distance protection. We propose a closed-loop signals directly from a copper connection is also set up as a
test model to perform benchmark line distance protection tests by reference. An Ethernet packet loss condition is then introduced
comparing the protection performance of relays that receive to the process bus communications system by a network traffic
analog signals via traditional copper wiring with relays that corruptor. This paper presents the results of this test.
receive analog signals via SV. We also discuss a test of the effects
of Ethernet packet loss on line distance protection and present the
II. INTRODUCTION TO LINE PROTECTION USING
results.
A PROCESS BUS NETWORK

I. INTRODUCTION In IEC 61850 digital secondary systems, the traditional

copper connections used to pass control signals, breaker and
Utilities invest in digital secondary systems (also known as
disconnect statuses, and other inputs are replaced with fiber
“digital substations”) to take advantage of benefits such as
connections that transmit GOOSE messages [1]. The copper
personnel safety and cost savings in copper cables and
installation time. IEC 61850 is the most common substation conductors that connect CTs or VTs to relays are replaced with
Ethernet-based SV messages over fiber optics [2] [3].
automation standard that these systems are based on; it specifies
Measurement samples are embedded in these messages and
Generic Object-Oriented Substation Event (GOOSE) and
published to subscribing SV relays via a process bus network.
Sampled Values (SV) messages for protection and control.
The unit that samples and converts analog measurements to SV
Though different from traditional systems, digital secondary
messages is an MU. A process bus network is a logical or
systems demand the same levels of protection and control
physical network that connects MUs and subscribing SV relays
system reliability and security as their traditional counterparts.
within a substation.
When designing a new digital secondary system, protection
engineers are often curious about the settings adjustments Fig. 1 shows a line protection scheme with MUs connected
to a process bus network, which distributes SV and GOOSE
necessary to implement SV-based systems in protective
messages to the SV relays. The process bus can be composed
relaying applications. One major area to consider for these SV-
of a fleet of switches, or it can be a logically segregated network
based systems is the challenges that can be introduced by the
in a substation communications network. The relays are also
communications network. Common network issues such as
connected to a station bus network, which in turn connects them
bandwidth limitations, latency, and packet loss must be
with supervisory control and data acquisition (SCADA)
considered. An SV-based line protection scheme must take
systems or human-machine interfaces.
 2

Fig. 2. Line differential protection scheme with a process bus network

III. COMMUNICATIONS CONDITIONS TO CONSIDER WITH

Fig. 1. Line distance protection scheme with a process bus network REMOTE DATA ACQUISITION VIA SV

A process bus network can be built using different network Fiber-optic cabling and network switches in digital
switch technologies. The most common process bus switches secondary systems replace the conventional copper cabling in
are Ethernet-managed switches. Engineers often use virtual traditional substations. As a result, an SV-based relay
local-area networks (VLANs) to segregate traffic and assign connected to a process bus can experience issues due to
traffic priorities to improve message delivery performance in bandwidth limitations, latency, or packet loss in the
network congestion situations. Software-defined networking communications channel.
for process bus traffic provides network path determinism and A. Bandwidth Limitations
cybersecurity [4]. Purpose-engineered network paths and On a process bus communications channel, there is often
monitoring for different types of traffic give system operators SV, GOOSE, and PTP traffic. An SV message compliant with
more confidence in and insight to the process bus network. IEC 61850-9-2LE [5] is approximately 150 bytes, which
SV-based protection and control applications require high- assumes an SV identifier (SVID) of 10 bytes and includes
accuracy time synchronization (such as IEEE 1588 Precision approximately 20 bytes of Ethernet frame overhead. When this
Time Protocol [PTP]) for MUs and relays, as shown in Fig. 1. SV message is published at 4.8 kHz, it consumes approximately
Each MU must provide an encoded sample number in each 5.760 Mbps of bandwidth. The two key factors that affect the
published SV message; these samples numbers are used to align remaining bandwidth are the size and publication rate of the
sampled analog measurements when the messages arrive at the GOOSE message. The size of a GOOSE message typically
relays. The process bus should be designed to impose the varies between hundreds of bytes to about 1,500 bytes,
minimum possible delay on the SV messages that it carries. depending on the number of binaries included. The GOOSE
SV-based line protection schemes depend on CT and VT publication rate varies between 2 ms and 1 minute, depending
signals from multiple locations and therefore subscribe to SV on the configurations of the MU and SV relay. PTP traffic is
messages from multiple MUs, as shown in Fig. 2. Sample time published once per second and includes approximately
coherency is critical to the security and reliability of the 480 bytes (when accounting for the four parts of PTP traffic:
protection algorithms, meaning that samples from multiple sync, announce, peer delay request, and peer delay response
MUs must be aligned. SV-based line differential protection messages). Due to the low publication rate, the total bandwidth
requires proper alignment of both MU data and local and needed for PTP is about 0.004 Mbps.
remote data.
 3

Engineers can calculate the network bandwidth not to exceed hundreds of microseconds. This network delay is
consumption based on the network design. Consider an caused by network switches and the optical signal transmission
example where the data from Table I is part of the network. In delay in fiber-optic cables.
this example, we assume that the network includes PTP traffic,
one SV stream, and 8 GOOSE messages containing 10 binaries
where one binary in each message changes state each second.
TABLE I
EXAMPLE NETWORK BANDWIDTH CONSUMPTION

Data SV GOOSE PTP

Bytes per
150 235 120
Message
Messages
4,800 40 4
per Second
Bandwidth
Consumption 5.760 0.075 0.004
(Mbps)
Includes one Fig. 3. Network delay and MU processing delay
Includes
8 GOOSE each of sync,
an SVID of These delays must be properly compensated for in SV
messages with announce, peer
10 bytes and
Assumptions 20 bytes of
10 binaries that delay request, relays, especially with analog signals coming from different
change state and peer delay MUs. Fig. 4 shows an example of two current measurements
Ethernet frame
once per second response
overhead
messages from two different MUs: one with the total SV channel delay
compensated for (IS) and one without SV channel delay
It is important to note that under normal system conditions, compensation (IR). In this example, the SV messages are
the bandwidth consumption is less than it is during an event published at 4.8 kHz and the total channel delay of the IR signal
condition, depending on the number of GOOSE messages and is 625 µs. The phase difference between IS and IR is
the rate at which they are published. However, when an event 13.5 degrees.
occurs on the system, the GOOSE messages containing state
changes are sent immediately and followed by their
retransmission sequences. This results in an increase in
bandwidth consumption.
With this concept expanded to an entire substation
installation, engineers should design an SV process bus
communications network taking the worst-case
communications bandwidth into consideration for both relay-
destined traffic and other normal network traffic. A common
way to manage other types of traffic on the network, such as SV
and GOOSE messages not subscribed to by a relay, is to use Fig. 4. Example of phase difference introduced by an uncompensated signal
traditional managed switches with VLANs. These VLANs One method to verify that an SV relay appropriately
segregate multicast and broadcast traffic to avoid unnecessary compensates for latency is for engineers to test a traditional
Ethernet message processing. Network traffic congestion is one relay and an SV relay in a side-by-side test. Each system needs
of the most common causes of high latency and jitter. An a common time source and needs to have common analog
incorrectly designed communications network could result in signals applied. An engineer can trigger a COMTRADE event
packet loss and, subsequently, a loss of protection. and compare the results. If the SV relay appropriately
B. Latency compensates for latency, then the COMTRADE data from both
Latency is the time it takes for an analog sample taken at the the SV relay and the traditional relay should have no phase
MU to arrive at an SV relay. This section discusses the total SV shift.
channel delay, which includes the MU processing delay and the C. Packet Loss
network delay [6], as depicted in Fig. 3. A packet loss condition occurs when an SV message
For line protection applications, the total SV channel delay published by an MU does not reach the subscribing SV relay.
should be actively monitored. The MU processing and network This may occur because of congestion on the process bus
delays can be calculated if the MU and the SV relay are network, a hardware failure in the system, or a bad fiber-optic
synchronized to a high-accuracy time source. The total MU cable. A sample counter indicating the time that the
processing delay for a protection application should not exceed measurement is taken within a one-second window is encoded
2 ms, according to IEC 61869-9 [3]. The network delay varies in the SV message. This sample counter increments from 0000
depending on the network architecture but is generally expected
 4

to 3,999 or 4,799, depending on the publication rate, and resets budget for network delays introduced by fiber optics and
at the top of every second. The relay leverages the sample Ethernet switches should be on the order of 100 µs.
counter information published in the SV message to identify This section discusses line protection basics, effects of
packet loss. communications-related issues on line protection, and
For SV messages, packet loss on the network causes analog recommendations on how to ensure the security and reliability
data loss at the SV relay. Fig. 5 shows the same IR signal from of SV-based line protection.
Fig. 4 but with five consecutive Ethernet messages lost. Packet
A. Line Differential Protection Overview
loss may also cause a reduced magnitude of filtered analog
quantities, which may prevent or delay protection operation. Differential protection operates on the sum of the current
entering and the current leaving a protected zone. The
differential current is proportional to the fault current for
internal faults and approaches zero for any other non-operating
(ideal) conditions. Differential relays calculate the differential
current based on instantaneous or phasor quantities. The
instantaneous differential current for a two-terminal
transmission line is defined as shown in (1).
i D= i L + i R (1)
where:
Fig. 5. Example of packet loss in addition to a phase difference introduced iD is the instantaneous differential current.
by an uncompensated signal
iL and iR are instantaneous local and remote currents
D. Other Considerations entering the protected zone as measured by relays at the
In addition to common network communications issues, SV- line terminals.
based systems require time synchronization between all the Local and remote relays use a communications channel to
MUs and relays. A loss of time synchronization can cause transfer their current measurements from one end of the line to
protection misoperation if an SV relay is not designed with the other, as shown in Fig. 6.
proper schemes to handle this loss or a time resynchronization
event. However, this paper focuses on communications-related
considerations and testing. For time synchronization-related
testing, refer to [6].
E. Impact of Communications Conditions
As demonstrated by the examples in this section, bandwidth
limitations, latency, and packet loss can severely impact signal
accuracy. An engineer should understand the bandwidth
limitation of the network and design for the worst-case Fig. 6. Line differential protection scheme with two terminals in SV-based
condition to prevent network congestion. SV-based relays substations
should measure the latency and account for it in protection
The percentage differential element compares an operating
applications. The relay should also be able to tolerate some
current and a restraining current against a user-defined
amount of packet loss. If security measures are not taken to
percentage threshold, K, and a minimum pickup threshold, KO,
compensate for these conditions, misoperations in power
as shown in Fig. 7. The operating current, IOP, is defined as
system protection can occur. Protection engineers must study
shown in (2).
the impact of communications conditions on protection
schemes before deploying an SV-based line protection system. IOP
= IL + IR (2)

IV. LINE PROTECTION AND THE EFFECTS OF where:

COMMUNICATIONS ISSUES IL and IR are the phasor local and remote currents
Protection engineers must consider the delays associated entering the protected zone as measured by relays at the
with SV communications and their impacts on overall two terminals.
protection speed. Engineers must also consider the impacts of The restraining current, IRT, is commonly defined as shown
SV communications delays on coordination timer settings, as in (3), (4), and (5), where k is a constant scaling factor. The
well as how a loss of SV can affect different protection relay operates when the system conditions result in the
elements. Again, an MU should have a maximum processing differential current plotting in the operating region of the
delay of 2 ms when used for protection [2]. The trip transfer characteristic.
time between the SV relay and the MU should be less than 3 ms
in protection applications, according to IEC 61850-5 [7]. The I RT k IL − IR
= (3)
 5

(
I RT k IL + IR
= ) (4) 1) Effects of Communications Issues on Differential
Protection in SV-Based Substations—Challenges
(
I RT = Max IL , IR ) (5)
and Recommendations
The loss of current measurements as a result of
communications issues can adversely impact line differential
protection [8]. As both local and remote currents are necessary
to determine if a fault exists on the system, loss of either type
of data affects the calculations in (1) through (5), as well as the
ratio of the local and remote currents. SV line differential relays
account for loss of current measurement conditions and take
appropriate action, such as blocking protection, to prevent a
misoperation.
Fig. 9 plots IOP versus IRT for a two-terminal line under
Fig. 7. Traditional percentage differential characteristic normal system conditions. Under nominal load, IRT = 2 per unit
(pu) and IOP = 0 pu. If the MU on one end of the line fails, it
Another common line differential protection method is
causes either the local or remote relay to stop receiving current
Alpha Plane differential. This method does not compare the
data. If the SV relay did not account for the loss of current
restraining current and operating current against a percentage
condition, the impact of this on percentage differential
threshold. Instead, it plots the ratio of the local current and
protection would be as shown in Fig. 9. Depending on the
remote current (IL / IR), which is a complex number, on an
minimum pickup settings, this could cause the relay to trip if
Alpha Plane, as shown in Fig. 8. The plane is defined with
the element is not supervised properly.
restraining and operating regions.

Fig. 9. Effect of losing SV data in a percentage differential application

If the SV relay did not account for current data loss in Alpha
Plane differential protection, the operating point (IL / IR) would
Fig. 8. Traditional Alpha Plane operating characteristic move from 1∠180 degrees (a restraining condition) toward the
The following are general principles for Alpha Plane current origin of the Alpha Plane, as shown in Fig. 10. Since the Alpha
differential protection: Plane operating region includes the origin to accommodate for
• Operating and restraining quantities are calculated current outfeed conditions, this would cause an improper trip
using the local current and the aligned remote currents operation if the protection element is not supervised properly.
for an n-terminal system. A generalized Alpha Plane
calculation can condense the n-terminal system into an
equivalent two-terminal system, which yields the same
operating and restraining quantities [8]. The
equivalent local and remote currents are then used in
Alpha Plane line differential protection.
• The presence of a substantial operating quantity is
used as a supervisory check.
• The complex ratio of the equivalent local current to
the equivalent remote current is plotted on an Alpha
Plane. For a relay to operate, the operating point must
be located in the tripping region, and the magnitude of
the operating quantity must exceed a threshold.
Fig. 10. Effect of losing SV data in an Alpha Plane line differential
application
 6

Line differential relays using SV data acquisition must block In this case, the SV relay data acquisition path delay includes
the differential element upon a loss of the current data used for the total SV channel delay. The total SV channel delay for the
differential protection. The relay must also send a blocking SV relay is typically a few milliseconds more than that of the
signal to remote relays over the line differential traditional relay with copper wiring. Furthermore, traditional
communications channel so that all relays in the zone of line differential data alignment does not compensate for such
protection remain secure. Similar logic exists in a line delays.
differential relay with traditional data acquisition (for dealing Fig. 12 illustrates the effect of a 1.5 ms data acquisition
with internal diagnostic errors, line differential channel delay mismatch for a 60 Hz system. During a load condition,
watchdog errors, and so on). Protection engineers should the IL / IR ratio ideally plots at 1∠180 degrees within the
include the SV channel status as another input to the existing restraining region. With an uncompensated data acquisition
blocking logic. delay mismatch of 1.5 ms, the operating point on the Alpha
Plane moves from the negative real axis by approximately
2) Line Differential Protection in Hybrid Installations
With Traditional and SV Relays—Challenges 30 electrical degrees. The operating point stays within the
and Recommendations restraining region, but there is a smaller margin for other errors,
Line differential relays must properly align local and remote such as CT errors and line differential channel asymmetry. This
current measurements. Data alignment compensation must be illustration emphasizes the need to take data acquisition delays
implemented for communicating current measurements from into consideration.
MUs to SV relays, as well as to compensate for the channel
delay between the local and remote relays. Channel-based data
alignment may use the ping-pong method so that the line
current differential channel delay can be measured without the
need for external time sources, assuming that the
communications channel delays (local-to-remote and remote-
to-local) are symmetrical. For asymmetrical communications
channels, high-accuracy external time sources can be used to
align local and remote current data.
Typically, the data alignment design assumes that the delays
between the primary equipment and the relays are identical
within each substation and that only the line current differential
communications channels need to be compensated for.
Fig. 12. Effect of uncompensated data acquisition delay mismatch on the
However, this assumption is sound only when the data Alpha Plane during normal load
acquisition systems at both ends of the line are the same. SV
line differential relays take data alignment conditions into For hybrid line differential applications using traditional and
consideration and compensate to prevent a misoperation. SV relays, it is undesirable to reduce protection sensitivity to
The case in Fig. 11 shows two relays in a line differential account for data misalignment resulting from mismatched data
protection scheme that are identical except for their data acquisition delays. Instead, the SV relay should compensate for
acquisition systems. One relay has a traditional data acquisition the total SV channel delay when time-stamping the line current
system with internal instrument transformers, and the other differential data packets. This guarantees proper data alignment
relay subscribes to SV data published by an MU. in hybrid applications and minimizes the impact to existing
non-SV differential relays.
B. Line Distance Protection Overview
A distance element uses current and voltage data and the
complex impedance plane (the R-X plane) to analyze distance
element operation. A typical mho element operating
characteristic is plotted in Fig. 13.

Fig. 11. Hybrid line differential installation with a traditional relay and an
SV relay
 7

user-programmable logic to supervise the negative-sequence

and zero-sequence overcurrent elements.
2) Effects of SV Voltage Data Loss on
Distance Protection and Recommendations
When SV voltage data are lost, the resulting effect on the
relay can be similar to a loss-of-potential (LOP) condition,
depending on whether current data are also lost. With voltage
measurements included in an SV message, SV packet loss will
result in the removal of all three voltage phases rather than
resulting in the typical single-phase voltage loss associated with
a true LOP condition. Modern protective relays typically alarm
upon detecting an LOP condition and disable distance elements
and directional elements that rely on voltage to operate. While
the protection scheme disabling is desirable when voltage is
missing due to SV data loss, an LOP alarm assertion is
Fig. 13. Impedance plane representation of distance element operation undesirable because it may falsely indicate a blown VT fuse.
Other relay elements, such as undervoltage, overvoltage,
The apparent or measured impedance Z is calculated and power elements, are also susceptible to unintended
according to (6), where the voltage and current measured at the operation as a result of SV data loss. Overvoltage elements are
relay are respectively V and I . Equation (6) is simplified to only at risk if they use negative-sequence or zero-sequence
illustrate the basic concept of line distance protection. voltages as operating quantities. Relays using voltage-based
V frequency measurement and tracking may also briefly
Z= (6) experience spikes in the measured frequency upon SV voltage
I
data loss or restoration. Power elements are similarly affected
During normal loading conditions, the measured impedance depending on the quantities used to calculate power and if either
seen by the relay is determined by the load flow. Typically, the the SV current or voltage data are lost.
impedance plots close to the real axis of the impedance plane. Upon loss of SV voltage data, the relay should block the
When a bolted fault occurs on a protected line, the measured distance and directional elements like it would during an LOP
impedance rapidly changes from Z = ZLOAD to Z = ZFAULT , condition, but it should not issue an LOP alarm. This secures
where ZFAULT is equal to the positive-sequence impedance of voltage-based protection but does not falsely indicate a blown
VT fuse. SV-specific channel monitoring Boolean quantities
the line between the relay location and the fault location.
are sufficient to indicate SV communications-related problems.
1) Effects of SV Current Data Loss on Protection engineers setting the relay should plan for delays in
Distance Protection and Recommendations protection that may result from a data loss condition. They
Modern protective relays have open-phase and open-pole should also use SV voltage channel status Boolean quantities in
detection logic, which supervise protection elements when the user-programmable logic to supervise undervoltage elements,
relay measures very low current. The logical assertions of these as well as overvoltage elements using negative-sequence or
elements may be stored in a Sequential Events Recorder for zero-sequence operating quantities. Upon loss or restoration of
analysis purposes. Loss of SV current data can result in a false SV voltage data, the relay should freeze the measured
open-phase declaration. frequency until the relay internal filters have time to stabilize (a
Loss of current data due to loss of SV can compromise the few cycles). Boolean quantities should indicate the status of the
dependability of distance elements, directional elements, and frequency measurements at all times. Similarly, for a power
breaker failure protection. Negative-sequence and zero- element, a protection engineer setting the relay should use both
sequence overcurrent elements may experience false operations the SV current and voltage channel status Boolean quantities to
due to the loss of a single current phase or due to filtered supervise the element.
transients when all three current phases are lost.
3) Effects of SV Data Loss on Communications-Assisted
To address these conditions, the SV relay should freeze the
Protection Schemes and Recommendations
open-phase or open-pole detection logic when SV current data
In general, the combined MU and process bus network
are lost. It should then hold the freeze for an extra few cycles to
delays result in an overall data acquisition delay of a few
allow the relay filters to stabilize after SV data are restored.
milliseconds. This affects nearly all protection elements.
This avoids nuisance operation of open-phase and open-pole
Communications-assisted distance protection schemes may be
elements. The overcurrent elements associated with breaker
either permissive schemes (such as permissive overreaching
failure logic can be handled in a similar way to partially
transfer trip) or blocking schemes (such as directional
preserve the dependability of those elements. Protection
comparison blocking). SV data loss inherently compromises
engineers setting the relay should plan for additional delays in
the dependability of permissive schemes and the security of
protection that may result from a data loss condition. They
blocking schemes. In addition, the total SV network delay
should also use SV current channel status Boolean quantities in
 8

needs to be accounted for when setting coordination timers used For this test, we placed the relay at the local line terminal.
in communications-assisted distance protection. We set up the mho element to protect the transmission line and
set the mho Zone 1 distance element with an 80 percent reach.
C. General Recommendations for Diagnostic and Security
The test simulated the faults at 50 percent of the line. There was
Measures for SV Subscriber Relays
no pickup time delay in the mho Zone 1 protection element.
SV subscriber relays should generate Boolean quantities
indicating the health of the signals they receive via SV. The B. Closed-Loop Benchmark Test—Traditional Distance
relays can use these quantities internally to supervise protection Relay vs. SV Distance Scheme
elements, either in a hard-coded fashion or via supervisory user- The purpose of this closed-loop benchmark test model is to
programmable logic equations. Engineers should also set relays measure the protection performance of an SV distance scheme
to ensure proper operation in case of transient data loss. It is against a traditional distance relay in the same protection
also desirable for the SV relay to be able to ride through the loss application. The SV distance scheme includes the combination
of one (or a few) packets by interpolating data internally. In of the MU and the SV relay. This provides baseline data for the
such a case, protection operation continues uninterrupted, and trip delays that can be expected in an SV-based substation
the relay issues no alarms. Long data outages should result in compared to a traditional substation. The test setup is shown in
selective blocking of protection elements, and the relay should Fig. 15. The real-time simulation tool provides the same
issue a communications alarm. voltage and current inputs to the MU and Distance Relay 2.
Distance Relays 1 and 2 use the same protection settings.
V. CLOSED-LOOP SV-BASED LINE DISTANCE Distance Relay 2 has a TRIP output directly connected to the
PROTECTION SIMULATION AND TESTING simulation tool, while a GOOSE message from Distance
A. Line Distance Protection Model Relay 1 drives the TRIP output on the MU.
To examine the concepts presented in this paper, we created
a test to benchmark SV protection performance and observe the
effects of communications-related issues. We modeled the
power system represented in Fig. 14 with a real-time simulation
tool. This power system comprises two power sources and one
transmission line where m is the distance to the fault. Table II
shows the system frequency, voltage, apparent power, source
impedance ratio (SIR) and impedance data.

Fig. 15. Benchmark test setup

The protection tripping time is the time from when the fault
Fig. 14. Power system simulation model
is applied to the time when the power system simulation tool
detects the TRIP output assertion. For this test, the simulation
TABLE II tool applied the same fault 1,000 times in both systems. This
POWER SYSTEM DATA approach allowed for an estimation of the average delay of the
Data Quantity Value SV-based distance scheme versus that of the traditional distance
relay. Fig. 16 shows the protection tripping times measured by
Frequency Fbase 60 Hz
the power system simulation tool.
Voltage Vbase(L-L) 230 kV
Apparent power Sbase 100 MVA
SIR – 0.5
Positive-sequence
ZL+ 160∠82° Ω (primary)
line impedance
Zero-sequence line impedance ZL0 480∠76° Ω (primary)
 9

trigger a fault and initiate a packet loss condition during the

fault condition to represent a worst-case scenario.
Fig. 18 displays the raw current waveforms and the filtered
current magnitudes from the two distance relays. Due to the loss
of more than three consecutive packets, the SV relay blocked
the distance logic calculation for the duration of the data loss.
Upon restoration of the SV data, the elements remain blocked
for an additional cycle to allow the relay data acquisition
elements to reach the steady state once again before declaring a
trip.

Fig. 16. Benchmark test results (performance comparison of SV distance

scheme versus traditional distance relay)

Table III displays the trip time statistics between the two
systems. In this example, the MU processing delay was less
than 1 ms. The remaining time is what it takes for Distance
Relay 1 to create and send a GOOSE message to the MU and
for the MU to process that message and trip the output.
Therefore, the GOOSE message processing time in the SV
distance protection scheme results in the majority of the delay
between the Distance Relay 2 trip and the MU trip.
TABLE III
BENCHMARK TEST TRIP TIMES

Average Maximum Minimum

Equipment Trip Time Trip Time Trip Time Fig. 17. Test setup with network traffic corruptor
(ms) (ms) (ms)
MU 20.48 22.73 17.50
Distance Relay 2 13.73 15.01 11.52

The benchmark time is the time difference between the TRIP

output assertion of the traditional relay and that of the MU in
the SV distance protection scheme. The average benchmark
time in this example was 6.75 ms. The benchmark time is a
useful reference during new product qualification processes to
understand the impact that SV-based protection has on tripping
times and to determine what settings need to be adjusted. Long
tripping times for line protection may affect power system
stability and should be carefully considered in stability Fig. 18. Packet loss results
simulations. Additionally, at commissioning testing,
benchmark results provide engineers with a baseline reference VI. CONCLUSION
to help determine the full impact of the process bus network
Key communications conditions to consider when
delay in an SV-based substation. The full delay introduced by
implementing an SV-based protection scheme are bandwidth
the process bus network also includes the processing delay of
limitations, latency, and packet loss. For bandwidth limitations,
each Ethernet switch the data passes through between the MU
it is important to calculate the worst-case bandwidth
and the distance relay. The Ethernet switch delay varies by
requirements and design the communications network to
manufacturer and technology, but the delay should be on the
accommodate this, thereby avoiding network congestion.
order of 100 µs for protection applications.
Latency resulting from the SV channel network delay can cause
C. Abnormal SV Communications Issue Simulation— a phase shift in the signal received at the SV relay. While
Packet Loss benchmark testing provides a baseline for the SV channel delay,
Fig. 17 displays the same test setup but with a loss-of-SV Ethernet network switches introduce additional delays.
condition introduced to the SV line relay. The network traffic Protection engineers should test for the total SV channel
corruptor intercepts packets and was configured to delete 40 of network delay during substation commissioning. An SV relay
every 4,800 packets. That equates to an 8 ms packet loss can compensate for this delay to an extent, but the system
duration every 1 second. The test setup was configured to should be designed to have less than 3 ms of total SV channel
 10

network delay. Finally, packet loss due to network congestion [7] IEC 61850-5, Communication Networks and Systems for Power Utility
Automation – Part 5: Communication Requirements for Functions and
or failed equipment can cause protection reliability issues.
Device Models, 2013.
While an SV relay can alarm for packet loss, a protection [8] H. Miller, J. Burger, N. Fischer, and B. Kasztenny, “Modern Line
engineer must monitor for this alarm and address the root cause Current Differential Protection Solutions,” proceedings of 63rd Annual
of the packet loss to ensure fast and reliable SV-based Conference for Protective Relay Engineers, College Station, TX,
protection schemes. March 2010.
SV-based line protection schemes consume SV messages
from multiple MUs, and the schemes are impacted by VIII. BIOGRAPHIES
communications conditions in several ways. In all cases, it is Steven Chase received his BS degree in electrical engineering from Arizona
State University in 2008 and his MS degree in electrical engineering in 2009.
important to monitor the SV channel status to make the He worked for two years as a substation design intern at Salt River Project, an
appropriate protection blocking or alarming decisions. For line Arizona water and power utility. He joined Schweitzer Engineering
differential protection, it is important for protection engineers Laboratories, Inc. in 2010, where he works as a lead power engineer in the
to use SV channel status as an input to existing blocking logic. research and development division. He is currently a registered professional
engineer in the state of Washington and a member of IEEE.
For line distance protection, the relay must discern between an
LOP condition and a loss of SV data. In both an LOP or a loss Erin Jessup received her BSEE and MEEE degrees from the University of
of SV data scenario, the distance element should be disabled, Idaho in 2005 and 2010. From 2005 to 2008, she worked as a product engineer
but it is important to not falsely declare a blown VT fuse. Other at Micron Technology in Boise, Idaho. Since 2010, she has worked at
Schweitzer Engineering Laboratories, Inc. in Pullman, Washington, and
common protection elements used in distance protection relays currently manages the distribution engineering department in R&D. In her
(such as breaker failure, undervoltage, overvoltage, power, and current role, she is responsible for supporting a wide variety of protective
directional elements) should be supervised with SV channel relays, as well as leading new product development. Erin is a registered
professional engineer in the state of Idaho and a member of IEEE.
status to make sure the data used in calculating trip decisions
are valid. SV relays include additional security measures to Mauricio Silveira is an electrical engineer with a BS earned from Sao Paulo
temporarily block logic that uses filtered data and allow signals State University in 2013. Since 2014, he has been with Schweitzer Engineering
to stabilize after SV data are restored. Laboratories, Inc. (SEL), where he has held positions in SEL Engineering
Services, Inc. (SEL ES), Sales and Customer Service, and R&D. He is currently
We provide a closed-loop SV-based benchmark test an integration and automation engineer. His work includes power system
example in this paper to determine the baseline SV channel modeling, cybersecurity assessment, and network design for critical
delay and the effects the delay has on distance protection and infrastructures.
tripping times compared with those of a traditional distance
Jiawei Dong received her bachelor’s of applied science in electrical
relay. Engineers can use this test model during initial product engineering from the University of British Columbia in 2016, and she earned
qualification to understand the impact that SV-based protection an MS in electrical engineering from the University of Idaho in 2018. She
will have on tripping times and to help them determine what joined Schweitzer Engineering Laboratories, Inc. in 2018, where she works as
relay settings need to be adjusted. Additionally, at an associate product engineer in the R&D division.
commissioning testing, this information provides engineers a
Qiaoyin Yang received her BS in electromechanical engineering from
reference to help determine the full impact of the process bus Guangdong University of Technology in 2010 and an MS in aerospace
network delay in an SV-based substation. engineering from North Carolina State University in 2012. Qiaoyin worked for
Schweitzer Engineering Laboratories, Inc. as a lead integration and automation
engineer in Pullman, Washington, from 2012 to 2019. She is a registered
VII. REFERENCES professional engineer, and she is a member of various working groups in the
[1] IEC 61850-8-1, Communication Networks and Systems for Power IEEE Power System Relaying and Control (PSRC) committee and The
Utility Automation – Part 8-1: Specific Communication Service International Council on Large Electric Systems (CIGRE) study committee B5.
Mapping (SCSM) – Mappings to MMS (ISO 9506-1 and ISO 9506-2)
and to ISO/IEC 8802-3, 2011.
[2] IEC 61850-9-2, Communication Networks and Systems for Power
Utility Automation – Part 9-2: Specific Communication Service
Mapping (SCSM) – Sampled Values Over ISO/IEC 8802-3, 2011.
[3] IEC 61869-9, Instrument Transformers – Part 9: Digital Interface for
Instrument Transformers, 2016.
[4] Q. Yang and R. Smith, “Improve Protection Communications Network
Reliability Through Software-Defined Process Bus,” proceedings of the
Grid of the Future Symposium, Reston, VA, October 2018.
[5] UCA International Users Group, “Implementation Guideline for Digital
Interface to Instrument Transformers Using IEC 61850-9-2,” July 2004.
Available: http://iec61850.ucaiug.org/Implementation%20Guidelines/
DigIF_spec_9-2LE_R2-1_040707-CB.pdf.
[6] Q. Yang, D. Keckalo, D. Dolezilek, and E. Cenzon, “Testing IEC 61850
Merging Units,” proceedings of the 44th Annual Western Protective
Relay Conference, Spokane, WA, October 2017.
© 2019 by Schweitzer Engineering Laboratories, Inc.
All rights reserved.
20190307 • TP6898-01