0% found this document useful (0 votes)
39 views10 pages

7071 LineCurrent RC 20230307 Web2

This document discusses challenges with using line current differential protection in systems with inverter-based resources and provides solutions to improve dependability and security. It presents guidelines for settings that account for the major sources of error from channel asymmetry and current transformer saturation to significantly enhance performance.

Uploaded by

baskaranjay5502
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
39 views10 pages

7071 LineCurrent RC 20230307 Web2

This document discusses challenges with using line current differential protection in systems with inverter-based resources and provides solutions to improve dependability and security. It presents guidelines for settings that account for the major sources of error from channel asymmetry and current transformer saturation to significantly enhance performance.

Uploaded by

baskaranjay5502
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 10

Line Current Differential Protection in Systems

With Inverter-Based Resources—Challenges


and Solutions

Ritwik Chowdhury, Ryan McDaniel, and Normann Fischer


Schweitzer Engineering Laboratories, Inc.

Presented at the
76th Annual Georgia Tech Protective Relaying Conference
Atlanta, Georgia
May 3–5, 2023

Previously presented at the


76th Annual Conference for Protective Relay Engineers, March 2023

Originally presented at the


49th Annual Western Protective Relay Conference, October 2022
1

Line Current Differential Protection in Systems With


Inverter-Based Resources—Challenges and Solutions
Ritwik Chowdhury, Ryan McDaniel, and Normann Fischer, Schweitzer Engineering Laboratories, Inc.

Abstract—Line current differential (87L) elements are popular Typical channels have asymmetries below 2 milliseconds;
for line protection and can provide excellent security and however, in rare cases, channels with multiplexers can have
dependability. In systems with inverter-based resources (IBRs), asymmetries as high as 5 milliseconds if not synchronized using
the elements may exhibit reduced dependability and, in cases when
not applied properly, degraded security. an external reference [3].
This paper presents challenges and solutions to 87L
applications in systems with IBRs. The solutions consist of settings
guidelines that can be used to significantly improve dependability
and security in such applications. The solutions may be
generalized to other challenging applications, such as series-
compensated lines with a possibility of current inversion, evolving
fault scenarios, or single-phase tripping applications.

I. INTRODUCTION
Line current differential (87L) elements are popular for line
protection due to their immunity to issues that challenge Fig. 1. Circuit relevant for 87L element
distance protection, such as power swings, load encroachment, TABLE I
uncertainty of line parameters, zero-sequence mutual coupling, IMPACT OF ERRORS ON 87L ELEMENT SECURITY AND DEPENDABILITY
and requirements for short-circuit studies (such as infeed and Source of Error Security Dependability
outfeed). As a tradeoff, they require a reliable communications
Channel asymmetry Major Moderate
channel and do not provide remote backup.
The 87L elements are an excellent choice in systems with CT saturation Major Minor
(during transients) (can misoperate) (can delay tripping)
inverter-based resources (IBRs) [1]. This is particularly due to
the numerous challenges faced by other line protection Line-charging Minor with compensation,
Minor
elements in these systems that can lead to reduced current otherwise can be moderate
dependability [2]. Steady-state CT and
Minor Minor
However, 87L relays can also face challenges in systems relay errors
with IBRs (as discussed in Section III). This paper presents System
No effect Moderate
solutions, primarily settings modifications, to improve the 87L nonhomogeneity
element performance in systems with IBRs (as discussed in
The asymmetry in the channel—or other causes of data
Section IV). The solutions characterize the major sources of
alignment errors, such as a time-source inaccuracy—translates
error for the 87L element, namely, channel asymmetry and
to an angular error for the 87L element but does not impact the
current transformer (CT) saturation. An application example
magnitude. The alpha plane (87AP) is developed to take
showing how to improve the reliability of the 87L element is
advantage of this observation and can be represented by the
included in Section V. The guidance is not limited to systems
logic of Fig. 2, with the radius and angle settings illustrated in
with IBRs and can be generalized to other applications.
Fig. 3, where IL and IR are the current phasors at the local and
remote line terminals or are calculated equivalents [4].
II. 87L ELEMENT OVERVIEW
A. 87L Element Requirements and the Alpha Plane
The circuit relevant to the 87L element is shown in Fig. 1.
The impact of the different sources of error on the 87L element
performance is summarized in Table I. A major source of error
for the 87L element is the channel. Applications with direct
fiber communications channels have minimal asymmetry.
Fig. 2. Simplified logic for the 87L element using the alpha-plane principle
2

Im (IL / IR) TABLE II


COMMON SETTINGS AND DEFAULTS FOR THE 87L ELEMENT

Setting 87LP 87LG 87LQ


Pickup 1.20 pu 0.25 pu 0.25 pu
Operate
region
Radius 6 6 6

B Angle 195 degrees 195 degrees 195 degrees


C
Radius
(87LPR) A A B. 87L Security for CT Saturation
–1 Re (IL / IR) 1) External Fault Detector (87EFD)
Angle The 87L element security for CT saturation may be
(87LPA)
A: Channel asymmetry addressed by use of an external fault detector (87EFD) to allow
Restraint B: System nonhomogeneity
C: CT saturation the relay to have two operating modes, a sensitive mode (when
region
87EFD = 0), and a secure mode (when 87EFD = 1) [8]. While
the secure mode is required to address the major source of
transient error associated with CT saturation (see Table I), the
sensitive mode must only account for steady-state errors.
When there is a possibility of CT saturation, 87EFD asserts,
Fig. 3. Alpha-plane characteristic, settings, and associated errors and the relay uses secure mode settings. The alternating-current
(ac) 87EFD shown in Fig. 4 can assert due to a significant
The operating characteristic associated with the 87AP logic increase in restraint current (DIRT) with little corresponding
of Fig. 2 is commonly represented by Fig. 3, assuming the increase in operating current (DIOP). It can also assert if there
element pickup is satisfied. The 87L element is further divided is excess direct current (dc) in any of the zone currents. The dc
into three subelements: 87EFD primarily improves 87L element security when
• Phase element (87LP)—the 87LP element performs a energizing a transformer or reactor external to the 87L zone [9]
per-phase differential (i.e., 87LA, 87LB, and 87LC) to [10] [11].
detect all fault types: line-to-ground (LG), line-to-line
(LL), line-to-line-to-ground (LLG), and three-phase
(3P).
• Zero-sequence element (87LG)—LG faults can have
high fault resistance [5]. The 87LG element
complements the 87LP element by adding sensitivity
for ground faults. The 87LG element has also been
applied in tapped line applications where the tapped
transformer has a delta winding on the high-voltage Fig. 4. Simplified 87EFD logic used to switch 87L element to secure mode
side and presents an open-circuit in the zero-sequence
As noted previously, the sensitive settings of the relay are
network. This means that the current downstream of
typically left to the values shown in Table II. The secure mode
the tap, with possibly expensive equipment associated
settings are approximately 20 percent higher by default to add
with the additional line terminal, is not required by the
security to the 87AP element.
87LG element to remain secure for faults downstream
of the tapped transformer. 2) Sequence Element Security
• Negative-sequence element (87LQ)—the 87LQ The larger blocking region in secure mode is adequate to
element complements the 87LP element by adding secure the 87LP element. However, the sequence differential
sensitivity for all unbalanced faults. In relation to the elements (i.e., 87LG and 87LQ) may not inherently have a
87LG element, the 87LQ element primarily adds restraint, for instance, during a 3P fault. In relays without an
sensitivity to LL/LLG faults with resistance. While 87EFD, they are blocked by the presence of dc and second
multiphase faults are typically considered to have harmonics [7] [12]. In relays with an 87EFD, their restraint is
small fault resistance, field events have shown that boosted by the fundamental phase currents and harmonics, so
this is not always the case [5] [6]. they are permitted to operate using secure mode settings, hence
The 87AP element settings correspond to the restraint region remaining dependable [8].
shown in Fig. 3. Most utilities use the default values shown in The sequence differential elements may lose dependability
Table II, where the settings have a per-unit (pu) base of the CT for internal faults with CT saturation, but since their objective
primary current rating. The alpha-plane operating principle has is to complement the 87LP element by detecting low-current
been applied to protect thousands of lines for over 20 years [7]. internal faults, this has historically not been an issue.
3

3) Settings Philosophy for Relays With the 87EFD approach is used in Section IV to provide settings guidance for
It is worth noting that the secure mode settings of a relay relays applying the 87AP element.
with the 87EFD can reuse the values shown in Table II [8],
since a relay without an 87EFD has used these settings and has III. PERFORMANCE CHALLENGES IN SYSTEMS WITH IBRS
demonstrated good performance for CT saturation for over
A. Dependability Issue
20 years [7].
The sensitive settings then, due to availability of the 87EFD, An investigation into the dependability challenges of 87L
can be reduced to only account for steady-state errors. This is relays is performed using a simulation of the power system near
typically not done in practice. Using settings biased towards an approximately 1,700 MVA Type 4 wind farm and a
security while operating in the sensitive mode reduces simulation of the relay algorithm [13]. To better understand the
dependability in systems with IBRs (as discussed in issue, the authors of this paper contacted the authors of [13].
Section III.A) or other systems (as discussed in Section V.B). The simulated event from [13] is played back through relay
In Section IV.B, we provide settings guidance to improve the hardware to observe the event record of Fig. 6.
87L element dependability. The event of Fig. 6 shows the three measured phase currents
at the two line terminals, with the currents at one line terminal
C. Alpha-Plane vs. Percentage-Restrained 87L Relays inverted, for a 15-ohm AG internal fault. The 87EFD in Fig. 6
The alpha-plane (87AP) and percentage-restrained (87PCT) asserts shortly after the fault. In this case, the 87EXFDO
relays both have a pickup setting that can be set similarly and dropout timer from Fig. 4 is shortened from 1 to 0.2 seconds to
must remain secure for the relatively constant sources of errors, show the event more clearly. For the entire duration that 87EFD
such as charging current. To remain secure for errors that is asserted, the 87L element uses secure settings and does not
increase with the current levels, such as channel asymmetry operate. This means that the 87L element may not trip for this
(linearly) and CT saturation (nonlinearly), the alpha plane internal fault for over 1 second.
offers two settings (radius and angle), unlike the percentage
differential (slope) [4].
The 87AP element radius is plotted as a function of the slope
of the 87PCT element for different values of blocking angles in
Fig. 5 for equivalent security. For this comparison, the slope
used by the 87PCT is based on a restraint that is the sum of the
individual current magnitudes forming the differential zone;
therefore, the slope can go up to a maximum value of
100 percent.

Fig. 6. 87L delayed operation (dependability issue) due to 87EFD assertion

The reason 87EFD asserts even though the fault is internal


is because of the significant change in currents on the unfaulted
Fig. 5. 87PCT slope for different 87AP blocking radius and angle settings
phases. The fault is internal to the A-phase, but it is external to
The angle setting is affected by the maximum level of the B-phase and C-phase. Since an 87EFD assertion on any of
asymmetry expected and can be calculated using (1). For the phases can switch the relay to secure mode, the element
instance, a worst-case asymmetry of 5 milliseconds is loses dependability for this scenario. In Section IV, we describe
associated with a blocking angle of 108 degrees in a 60 Hz how this possible loss of dependability is not unique to systems
system. with IBRs and discuss solutions to mitigate this issue.
Angle = Asymmetry • Frequency • 360 degrees (1) B. Security Issue
The default 87LP radius, and angle settings shown in In August 2019, 87L relays protecting a 0.2-mile, 138 kV
Table II correspond to an 87PCT slope setting of 90 percent, line near a 184 MVA Type 3 wind farm misoperated. The
which is heavily biased towards security. Settings guidance by associated field event report is shown in Fig. 7. The currents
relay manufacturers is typically not provided as a function of and voltages exhibit significant harmonic content. The 87LQ
system parameters but has been done in the past [10]. A similar element is set sensitively with a pickup of 0.10 pu (the
minimum possible setting of the relay); since the operate
4

current magnitude (87IQDIFM) exceeds this value, the relay • CT remanence up to 80 percent is tested and often
trips without any fault present. considered to be a good limit [10].
• The point-on-wave varies in 5-degree increments to
yield greater accuracy relative to the 30-degree
increments required by IEC 60255-187-1 [16].
Based on the test results, the 87AP element radius and angle
required for secure behavior due to CT saturation for external
faults is shown in Fig. 8. The secure settings shown in Fig. 8
are presented as a function of the total CT dimensioning
required (KTOT), which considers the transient dc offset and a
remanence level up to 80 percent [10] [18]. This KTOT value can
then be used to size CTs and set the relay, as shown in the
Section V.A example. The blocking radius shown in Fig. 8
corresponds to the inverse of the magnitude attenuation due to
CT saturation that the relay tolerates. The blocking radius acts
as a multiplier to the other lesser sources of errors. The blocking
angle shown in Fig. 8 corresponds to twice the angular error
associated with CT saturation. The blocking angle adds to other
sources of errors, such as channel asymmetry.

Fig. 7. 87LQ misoperates (security issue) near a Type 3 wind farm

The cause of the misoperation is attributed to the excessive


current harmonics that can manifest as an erroneous
negative-sequence differential current with a magnitude ripple.
IBRs have been known to produce excessive harmonics during
open-phase conditions or in weak grids [14] [15].

IV. SOLUTIONS TO IMPROVE DEPENDABILITY AND SECURITY


A. CT and Relay Requirements to Address CT Saturation
This section provides the CT and relay requirements of an
87L relay with the 87EFD [8] to establish limits to security.
This is important for Section V.A, where we consider an IBR
application. The details of the system and the method used to Fig. 8. 87AP blocking radius and angle requirements to accommodate CT
saturation for systems up to different X/R ratios
develop the requirements are summarized as follows:
• A 500 kV system is used with a similar topology as The KTOT required by the 87AP relays to perform according
the one shown later in Section V.A. The line length to the intended design is 4. Tuning settings based on CT
varies from 1 to 200 kilometers to test performance for dimensions and the X/R ratio of the worst-case external faults
short lines near generating plants with possibly high can add some dependability for internal faults even when
X/R ratios and for long lines with high-charging 87EFD is asserted, for instance, during the example in
current. Section III.A or during scenarios with evolving external-to-
• The X/R ratio of the fault currents varies up to a value internal faults. The settings adjustments are applicable to relays
of 100 to accommodate short-line or dual-breaker without an 87EFD [7] or the secure mode settings for relays
applications near generating plants. with an 87EFD [8]. The minimum KTOT by the 87AP relay with
• One terminal has dual breakers to test the an 87EFD [8] to ensure security for external faults is 2, which
requirements associated with an external fault at the corresponds to such severe saturation that the 87EFD does not
dual-breaker bus where unequal saturation of the dual- get a chance to assert and adapt the relay to secure mode.
breaker CTs can appear as a false differential current However, in rare applications that use CTs with a KTOT between
into the line [8]. 2 and 4, the 87AP element should be blocked when 87EFD
• Only one CT is saturated for the different faults to asserts, since no secure mode radius and angle settings can be
emulate worst-case mismatch and maximize the applied to guarantee secure operation for all possible external
erroneous differential current. faults.
• CT parameters are varied and use data from both An example simulated misoperation showing the security
limits of the 87AP relay applied at a dual-breaker terminal is
IEC 60255-187-1 and IEEE C57.13 [16] [17].
shown in Fig. 9. The KTOT for the saturated CTs is 3, and the
5

system X/R ratio is 100. The 87AP element is set with a secure intent is to ignore the possibly poor negative-sequence current
radius of 6 and an angle of 120 degrees, significantly lower than injected by the IBR.
what is expected based on the values shown in Fig. 8. In Fig. 9, SIBR
87EFD asserts even though CT saturation occurs within 87LQPSENS = 1.25 • pu (2)
4 milliseconds, but the secure mode settings are not sufficient 3 • VHV • ( CTR • I NOM )
to prevent a transient misoperation of the 87AP element. The secure pickup settings are based on the maximum
current of the IBR, which can be 1.10 to 1.30 pu higher [2] and
is represented by (3).
87LQPSECURE = 1.30 • 87LQPSENS pu (3)
The minimum values of 0.20 pu and 0.30 pu for the sensitive
and secure pickup settings, respectively, for the 87LQ and
87LG elements provide security for the steady-state CT and
relay errors. The pickup for the 87LP element is set as follows:
• Sensitive—the steady-state CT and relay errors are in
the order of 0.05 pu. Errors associated with charging
current compensation or for short lines when
uncompensated is typically less than 0.10 pu [19].
When adding a security margin for minor system
transients, a value of 0.30 pu is adequate.
• Secure—additional considerations can raise this
setting, such as significant transients associated with
energization of the protected line [7] [8] and
Fig. 9. Transient 87AP element misoperation case with heavy CT saturation
energization of a transformer or a reactor external to
B. Settings Guidelines the differential zone [11]. A value of 0.75 pu is
The 87AP element is set by considering the various sources usually adequate.
of errors, which are summarized in Table III for sensitive mode The 87AP radius is set with the following considerations:
(when 87EFD = 0) and Table IV for secure mode (when • Sensitive—mismatched relay and CT errors between
87EFD = 1). For secure mode, CT saturation is a major the two terminals are assumed to be as large as
contributor of errors. Based on field experience, line CTs are 15 percent. When adding a security margin to get a
typically dimensioned better than a KTOT of 6. A challenging total error of approximately 25 percent [11], the 87AP
application example that supports this claim is shared in element radius is 1.35.
Section V.A. • Secure—CT saturation is the main source of error.
TABLE III Considering a KTOT of 6, the associated radius based
SENSITIVE 87L SETTINGS GUIDELINES on Fig. 8 is 4. Adding a 25-percent margin for the
other sources of errors, the radius is 5.
Setting 87LP 87LG 87LQ
The 87AP angle is set with the following considerations:
Pickup 0.30 pu 0.20 pu 1.25 puC • Sensitive—the main source of error is channel
Radius 1.35 1.35 1.35 asymmetry, which is typically within 2 milliseconds
Angle 90 degrees A
90 degrees A
90 degreesA [3]. Assuming a conservative 3.5-millisecond
asymmetry for a 60 Hz system and an additional
TABLE IV 15-degree margin, we calculate a value of 90 degrees,
SECURE 87L SETTINGS GUIDELINES
as shown in (1).
Setting 87LP 87LG 87LQ • Secure—CT saturation adds an angular error, and a
Pickup 0.75 pu 0.30 pu 1.63 puC value of 80 degrees is adequate for a KTOT of 6. This
results in an overall angle of 170 degrees for security.
Radius 5.00B 5.00B 5.00B
According to our simulations and field events of
Angle 170 degreesA,B 170 degreesA,B 170 degreesA,B generator black-start and shunt reactor inrush [11], the
A
Adjust based on worst-case channel asymmetry (Section IV.B). angular error due to dc CT saturation when energizing
B
Adjust from a KTOT of 6 based on CT sizing guidelines (Fig. 8). a transformer or reactor external to the 87L zone
C
Adjust based on IBR rating and system parameters (Section IV.B).
typically does not exceed 25 degrees (or an 87AP
For the 87LQ element, the sensitive pickup is set based on element angle setting of 50 degrees). Hence, the
(2), which uses the MVA rating of the IBR (SIBR), the voltage values shown in Fig. 8 provide adequate security.
level of the line (VHV), the CT ratio (CTR), the relay nominal For relays without an 87EFD [7], only the secure settings
secondary current (INOM), and a margin of 25 percent. The should be used. If the application parameters are better (i.e.,
pickup is set based on similar principles from [2], in which the CTs are sized larger than a KTOT of 6 or the channel has a worst-
case asymmetry lower than 3.5 milliseconds), the user may
6

choose to use the default settings or use more dependable differential elements (87LG and 87LQ), which can still lose
settings. If the application parameters are worse, more secure dependability for an internal fault in these applications.
settings should be used.
D. Improved Performance of 87L Relay
C. Relay Design Improvements
1) Dependability
An 87EFD assertion reducing dependability, as discussed in The simulated event from Section III.A is played back to the
Section III.A, can occur in systems without IBRs as well. We relay with the settings from Table III and Table IV and with the
have seen such a response in conventional generator, default levels of channel asymmetry and CT saturation
transformer, and bus applications [20]. A simple example is the considered. Equations (2) and (3) are evaluated for the system
case of a generator step-up transformer (GSU) with an open with an IBR MVA rating (SIBR) of 1,700 MVA, a voltage rating
low-voltage generator circuit breaker (GCB) where the GSU (VHV) of 500 kV, and a CT ratio of 2000/5 (CTR = 400,
presents a low-impedance zero-sequence path for an LG fault INOM = 5 A). The resultant 87LQ sensitive and secure settings
on the high-voltage side [21]. The example is shown in Fig. 10 are 1.23 pu and 1.60 pu, respectively.
for an AG fault on the line. Before the fault, due to the open The relay event with the improved settings is shown in
GCB, there is no current; during the fault, pure zero-sequence Fig. 11. Even though 87EFD asserts, the relay trips on 87LP,
current is seen at the GSU terminals with the A-, B-, and 87LG, and 87LQ. The 87LQ element has delayed assertion,
C-phase CTs measuring the same current. The A-phase current which can happen due to the incoherent negative-sequence
flows from both line terminals to the fault, and the 87L element current injected by the IBR [1].
sees the internal fault on the A-phase; the B-phase and C-phase
87L elements see the current flow through the unfaulted phases
classifying it as an external fault, thereby asserting 87EFD.

Fig. 10. Pure zero-sequence current seen at transformer terminals for


AG fault

On differential relays that use only local measurements (e.g.,


generator, transformer, or bus relays), this scenario is not a
problem since the relays have fully phase-segregated
differential elements, including the external fault detector [10]
[20] [21]. In other words, an external fault detector assertion in
the B-phase or the C-phase does not reduce dependability in the
A-phase.
In 87L relays, the design may be slightly different due to the Fig. 11. Dependable operation of 87L elements despite 87EFD assertion
requirement of a communications channel. To provide the most 2) Security
flexibility and to support channels with minimal bandwidth For the field event of Section III.B where 87LQ misoperates,
(1 • 64 kbps) [22], the packet contains and transmits one 87EFD SIBR is 184 MVA, VHV is 138 kV, CTR is 400, and INOM is 5 A.
bit to the remote relay. Transmitting a single 87EFD bit adds Using (2) and (3), we calculate a sensitive and secure pickup of
security for an external fault behind a remote dual-breaker 0.48 pu and 0.63 pu, respectively. From Fig. 7, it is evident that
terminal [8], but it can reduce dependability of the other phases. a sensitive pickup setting of 0.48 pu can provide adequate
For 87L applications with conventional generation that is security for this event.
online, the change in unfaulted phase currents through the line
is not significant enough where 87EFD asserts due to the E. Sensitivity Comparison Between 87AP and 87PCT
significantly lower associated positive-sequence and negative- The relative sensitivities of the 87AP and 87PCT phase
sequence impedances. However, in systems with IBRs, since element (87LP) are evaluated by applying LG faults of various
they are weak sources, the phenomenon is expected to occur resistances on a 230 kV, 50-kilometer line. The line has one
more frequently, even when the IBR plant is online. Designing weak terminal with a positive-sequence source-to-line
the 87L element so that the 87EFD is phase-segregated and only impedance ratio (SIR1) of 5 and a zero-sequence source-to-line
secures the associated phase differential element restores impedance ratio (SIR0) of 0.5 representing the IBR and the
dependability. This requires communicating three bits, one for interconnection transformer. The other, strong terminal
each phase, in the packet. This does not benefit the sequence represents the grid and has an SIR1 and SIR0 of 1.
7

The 87AP element is set using the default sensitive settings V. APPLICATION CONSIDERATIONS
of Table III, whereas 87PCT is set with an equivalent slope
A. CT Guidance Applied in an IBR System Example
setting of 40 percent, according to Fig. 5. The sensitivity
comparison of the 87AP and 87PCT phase elements at full load The settings guidelines of Section IV.A associated with CT
is shown in Fig. 12, including the reference pickup of 0.30 pu, saturation are applied to a typical IBR tie-line example, shown
assuming no assertion of the 87EFD. The 87LP pickup, which in Fig. 14. A 138 kV, 20-kilometer tie line connects the IBR
is common to both 87AP and 87PCT, is shown as reference. No plant from the point of measurement (POM) to the three-
channel asymmetry or CT saturation is modeled for this breaker ring bus at the switching station labeled as the point of
comparison, but the elements are set by considering the errors interconnection (POI). All CTs have a ratio of 200, an internal
associated with the sensitive settings, as explained in CT resistance (RCT) of 0.5 ohm, and a burden (RB) of 1 ohm.
Section IV.B.

Fig. 12. 87AP and 87PCT phase element sensitivity comparison

It is evident from Fig. 12 that the 87AP element provides Fig. 14. Example of IBR plant connected to a three-ring breaker bus
greater sensitivity. This is because the 87AP element can trip if
Three-phase faults may occur at the location F1 and the
either the current magnitude or angle between the two line
location F2, and the scenarios are explained as follows.
terminals deviates (see Fig. 2 and Fig. 3). In these examples, the
• The maximum fault current seen by POM Relay R1,
current magnitude is sufficiently different, allowing the 87AP
as measured by CT1, is for a fault at F1 where the grid
element to provide greater sensitivity. In contrast, the relatively
contribution is 5 kA with an X/R of 10.
large slope of the 87PCT element makes no distinction between
magnitude and angle deviations. Having only the slope setting • The maximum fault current seen by POI Relay R2, as
to accommodate possibly large channel asymmetry limits measured by CT2 and CT3, is for a fault at F2 with
87PCT element sensitivity. CB4 open and the contribution mostly from the
Modern communications are typically quite good; and for system at 10 kA with an X/R of 20.
applications where the worst-case asymmetry is expected to be
small, the 87AP and the 87PCT elements can be set sensitively
to perform similarly, as shown in Fig. 13. Unlike 87PCT, the
87AP element sensitivity does not decrease significantly with
channel asymmetry, because the ratio of the current magnitudes
at the two line terminals during a fault is sufficiently different.
Channels with small asymmetry promote sensitive settings,
which benefit the applications described in Section V.B.

Fig. 13. Sensitivity comparison of 87AP and 87PCT phase element with
equivalent settings for different levels of channel asymmetry
8

The KTOT required by the relay to retain dependability in • Series-compensated lines—current inversion in some
high-security mode is 4. The saturation voltage (VSAT) of the systems is a possibility for an internal fault [23].
CT may be calculated based on (4), with the values for CT1 and • Single-phase tripping applications—increased
CT2/CT3 shown in (5) and (6), respectively. sensitivity from the 87LP element can improve phase
=VSAT K TOT • I F • ( R CT + R B ) (4) selectivity for high-resistance faults in these
applications.
 5, 000 A  • Scenarios with evolving external-to-internal faults—
VSAT _ CT1 4 • 
= =  • ( 0.5 Ω + 1 Ω ) 150 V (5) it is possible for 87L relays to have reduced
 200 
dependability when a fault evolves from an external to
 10, 000 A  an internal one due to the bigger secure mode restraint
VSAT _ CT23 4 • 
= =  • ( 0.5 Ω + 1 Ω ) 300 V (6) region.
 200 
The guidance may also be applied to other applications to
The minimum VSAT for CTs rated C100, C200, C400, and add sensitivity while maintaining adequate security. Having an
C800 with a 0.5-ohm RCT is 150, 250, 450, and 850 V 87L channel provides numerous other advantages, including
respectively [10]. For CT1, any of these CTs may be applied. improved fault-type identification and fault location due to
For CT2/CT3, since the worst-case VSAT of the application availability of currents from both line terminals. The relay may
is 300 V, the next highest CT class of C400 is preferable. The also use the low bandwidth (1 • 64 kbps) channel to provide
minimum KTOT required by the relay is 2 (as discussed in double-ended traveling-wave fault location, which provides
Section IV.A), which is satisfied by a C100 CT; however, in significantly higher accuracy than impedance-based fault
such cases, it is expected that the 87EFD bit is applied to block location techniques [8].
the 87AP element instead of elevating it to secure mode.
If we assume that C400 CTs are used for this application, VI. CONCLUSION
then the relay settings may be made more sensitive. For a C400
The 87L element is an excellent choice in systems with
with a VSAT of 450 V applied at both terminals, the effective
IBRs. When applied with conventional guidance, the 87L
KTOT may be calculated as follows:
element may lose dependability for ground faults in systems
• CT1 has a KTOT of 4 • (450 V/150 V) = 4 • 3 = 12.
with IBRs. This loss of dependability occurs when there is a
• CT2 and CT3 have a KTOT of 4 • (450 V/300 V) =
significant change in unfaulted phase currents due to the low-
4 • 1.5 = 6.
impedance zero-sequence path presented by the IBR
For this application, the worst-case X/R ratio is greater than
transformer, whereas the IBR acts as a weak positive-sequence
15 and the worst-case KTOT is 6. To accommodate worst-case
and negative-sequence source. The 87LQ element may lose
CT saturation, we obtain a radius setting of 4 and angle setting
security due to the harmonic content and poor negative-
of 80 degrees from Fig. 8 using the blue line representing
sequence injected by the IBR, as we showed using a field event.
X/R = 100. With a 25-percent margin for the radius and
This paper presented solutions to improve 87L element
90 degrees for angular errors due to channel asymmetry
dependability and security via improved application guidance
including a margin (as discussed in Section IV.B), the radius is
and relay design improvements. The security improvement
5 and the angle is 170 degrees.
consists of desensitizing the 87LQ element based on the IBR
If C800 CTs are applied at the dual-breaker POI terminal,
ratings and basic system data. The dependability improvements
the effective KTOT is 11.33. In that case, a secure mode radius
are achieved by characterizing the two potential major sources
of 4.6 and an angle of 150 degrees can be applied. It is evident
of errors for the 87L element, namely channel asymmetry and
from this example that CTs at the dual-breaker terminals have
CT saturation. The 87L element can also be designed to send
higher requirements than single-breaker terminals, because the
additional data in the communication packet by using the higher
worst-case external fault current and the associated X/R ratio is
bandwidth afforded by modern communications infrastructure.
not reduced by the line impedance. For this example, we use
The added dependability significantly benefits applications in
three-phase faults, but a similar calculation applies to ground
systems with IBRs, series-compensated lines with current
faults. IEC CT sizing and relay settings can be done in a similar
inversion, systems using single-phase tripping, and evolving
manner [10].
fault scenarios.
If the system fault current levels are lower due to higher
penetration of IBRs or for a weaker grid, the element can be set VII. ACKNOWLEDGEMENTS
more sensitively.
The authors thank Aboutaleb Haddadi and Evangelos
B. Application Dependability Improvements Farantatos of EPRI for sharing a simulated event used for this
The dependability improvements explained in this paper paper. The authors also thank Brian Smyth for sharing his
benefit many applications considerably, including the experience supporting 87L relays.
following:
• Systems with IBRs—as explained in Section III.A, it
is possible for the 87EFD to desensitize the 87L
element in such applications.
9

VIII. REFERENCES [20] A. B. Dehkordi, R. Chowdhury, N. Fischer, and D. Finney, “Generator


Protection Validation Testing Using a Real-Time Digital Simulator:
[1] R. Chowdhury and N. Fischer, “Transmission Line Protection for
Stator Winding Protection,” proceedings of the 48th Annual Western
Systems With Inverter-Based Resources – Part I: Problems,” IEEE
Protective Relaying Conference, Spokane, WA, October 2021.
Transactions on Power Delivery, IEEE, Vol. 36, No. 4, August 2021,
pp. 2,416–2,425. [21] R. Cole, R. Tuck, T. Solinsky, C. Sun, R. Chowdhury, A. Abd-Elkader,
and B. Matta, “Bus Differential Protection Upgrade for a 1,500 MVA
[2] R. Chowdhury and N. Fischer, “Transmission Line Protection for Nuclear Plant With Atypical Connections,” proceedings of the
Systems With Inverter-Based Resources – Part II: Solutions,” IEEE 75th Annual Conference for Protective Relay Engineers, College
Transactions on Power Delivery, IEEE, Vol. 36, No. 4, August 2021, Station, TX, March 2022.
pp. 2,426–2,433.
[22] IEEE Std C37.94-2017, IEEE Standard for N Times 64 kbps Optical
[3] IEEE Std C37.243-2015, Guide for Application of Digital Line Current Fiber Interfaces Between Teleprotection and Multiplexer Equipment.
Differential Relays Using Digital Communication.
[23] E. Bakie, C. Westhoff, N. Fischer, and J. Bell, “Voltage and Current
[4] D. Tziouvaras, H. Altuve, G. Benmouyal, and J. Roberts, “Line Inversion Challenges When Protecting Series-Compensated Lines – A
Differential Protection With an Enhanced Characteristic,” proceedings Case Study,” proceedings of the 42nd Annual Western Protective Relay
of the 3rd Mediterranean Conference on Power Generation, Conference, Spokane, WA, October 2015.
Transmission, Distribution, and Energy Conversion, Athens, Greece,
November 2002.
[5] C. Henville and R. Chowdhury, “Coordination of Resistive Reach of IX. BIOGRAPHIES
Phase and Ground Distance Elements,” proceedings of the 48th Annual Ritwik Chowdhury received his BS degree in engineering from the University
Western Protective Relay Conference, Spokane, WA, October 2021. of British Columbia and his MS degree in engineering from the University of
[6] D. Miller, “Three Phase Bus Fault With Fault Impedance,” proceedings Toronto. He joined Schweitzer Engineering Laboratories, Inc. (SEL) in 2012,
of the 42nd Annual Western Protective Relaying Conference, Spokane, where he is presently a senior engineer in research and development. Ritwik
WA, October 2015. holds 9 patents and has helped author 25 technical papers. He was recognized
as an exceptional reviewer for IEEE Transactions on Power Delivery for 2019
[7] SEL-311L Line Current Differential Protection and Automation System and 2021. He is the vice chair of the Protection and Control Practices
Instruction Manual. Available: selinc.com. Subcommittee (I-SC) of the IEEE PSRC Committee, the chair of two IEEE
[8] SEL-411L Advanced Line Differential Protection, Automation, and Standards Working Groups, and the recipient of the 2021 PSRC Outstanding
Control System Instruction Manual. Available: selinc.com. Young Engineer Award. Ritwik is a senior member of the IEEE and a registered
[9] B. Kasztenny and D. Finney, “Generator Protection and CT Saturation professional engineer in the province of Ontario.
Problems and Solutions,” proceedings of the 58th Annual Conference
for Protective Relay Engineers, College Station, TX, April 2005. Ryan McDaniel earned his BS in computer engineering from Ohio Northern
[10] R. Chowdhury, D. Finney, N. Fischer, and D. Taylor, “Determining CT University in 2002. In 1999, he was hired by American Electric Power (AEP)
Requirements for Generator and Transformer Protective Relays,” where he worked as a relay technician and a protection and control engineer. In
2005, he joined Schweitzer Engineering Laboratories, Inc. (SEL) and is
proceedings of the 46th Annual Western Protective Relay Conference,
currently a principal engineer. His responsibilities include providing
Spokane, WA, October 2019.
application support and technical training for protective relay users. Ryan is a
[11] R. Chowdhury, N. Fischer, D. Taylor, D. Caverly, and A. B. Dehkordi, registered professional engineer in the state of Illinois and a member of the
“A Fresh Look at Practical Shunt Reactor Protection,” proceedings of IEEE.
the 49th Annual Western Protective Relay Conference, Spokane, WA,
October 2022. Normann Fischer received a Higher Diploma in Technology, with honors,
[12] G. Benmouyal and T. Lee, “Securing Sequence-Current Differential from Technikon Witwatersrand, Johannesburg, South Africa, in 1988; a BSEE,
Elements,” proceedings of the 31st Annual Western Protective Relay with honors, from the University of Cape Town in 1993; an MSEE from the
Conference, Spokane, WA, October 2004, pp. 23–25. University of Idaho in 2005; and a PhD from the University of Idaho in 2014.
[13] A. Haddadi, E. Farantatos, I. Kocar, and U. Karaagac, “Impact of He joined Eskom as a protection technician in 1984 and was a senior design
Inverter Based Resources on System Protection,” Energies, Vol. 14, engineer in the Eskom protection design department for three years. He then
No. 4, February 2021, p. 1,050. joined IST Energy as a senior design engineer in 1996. In 1999, Normann
joined Schweitzer Engineering Laboratories, Inc., (SEL) where he is currently
[14] R. M. Moreno, J. A. Pomilio, L. C. Pereira da Silva, and S. P. Pimentel, a distinguished engineer in research and development. He was a registered
“Mitigation of Harmonic Distortion by Power Electronic Interface professional engineer in South Africa and a member of the South African
Connecting Distributed Generation Sources to a Weak Grid,” Brazilian Institute of Electrical Engineers. He is currently a senior member of IEEE and
Power Electronics Conference, Bonito-Mato Grosso do Sul, Brazil, a member of the American Society for Engineering Education (ASEE).
October 2009, pp. 41–48. Normann has authored over 60 technical and 10 transaction papers and holds
[15] J. Gahan, A. Valdez, B. Cockerham, R. Chowdhury, and J. Town, “Field over 20 patents related to electrical engineering and power system protection.
Experience With Open-Phase Testing at Sites With Inverter-Based
Resources,” proceedings of the 74th Annual Conference for Protective
Relay Engineers, Virtual Format, March 2021.
[16] IEC Std 60255-187-1-2021, Measuring Relays and Protection
Equipment – Part 187-1: Functional Requirements for Differential
Protection – Restrained and Unrestrained Differential Protection of
Motors, Generators, and Transformers.
[17] IEEE Std C57.13-2016, IEEE Standard Requirements for Instrument
Transformers, 2016, p. 38.
[18] IEC TR 61869-100:2017, Instrument Transformers – Part 100:
Guidance for Application of Current Transformers in Power System
Protection.
[19] A. Hargrave and G. Smelich, “Setting and Testing Line Charging
Current Compensation in the SEL-411L Relay,” SEL Application Guide © 2022, 2023 by Schweitzer Engineering Laboratories, Inc.
(AG2018-02), 2018. Available: selinc.com. All rights reserved.
20230307 • TP7071-01

You might also like