See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/3170396

Protective device coordination in an industrial power system with multiple

sources

Article in IEEE Transactions on Industry Applications · August 1997

DOI: 10.1109/28.605753 · Source: IEEE Xplore

CITATIONS READS

44 2,183

1 author:

P.E. Sutherland
Retired from GE and ABB
67 PUBLICATIONS 580 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Power Generation View project

Grounding View project

All content following this page was uploaded by P.E. Sutherland on 24 February 2015.

The user has requested enhancement of the downloaded file.

 Page 1 of 1
Lr

Robbv Villanueva

From: OAO <[email protected]>

To: <[email protected]>
Sent: Tuesday, October 31,2000 9:38AM
Subject: WWW ILL Request

[email protected] (OAO) sent the following ILL request:

Author: Sutherland,-Peter-E
Book Title: IEEE Transactions on Industry Applications
Journal Article: Protective device coordination in an industrial power system with multiple sources
Year: 1997
Volume: v 33 JulylAug
Pages: 1096 / 103
Location:
Rush: on
Needed By:
User Name: OAO
E-Mail address: elucas&oliat.espol.edu.ec
Phone #: 3 1 - 10-00
................................................................
Server Protocol: HTTP/l.1
Remote Host: (200.10.149.11)

11 15100
 IEEE TRANSACTIONS ON INDUSTRY APIIJCAlWNS. V O L 33. SO 4. KlY/AUCUST 1997

Protective Device Coordination in an Industrial

Power System with Multiple Sources
Peter E. Sutherland, Member, IEEE

Abstmct-Pmtective device coordination in an Industrial power

system with electric utility ties and multiple in-plant generators
presents challenges that cannot be resolved by the use of con-
ventional time-current curves. Relays in different locations will
detect greatly different currents during the same fault. This may
be resolved by the use of the conventional per-unit time-current
curve. However, many computer-aided tim-urrent curve pro-
grams do not support this technique. The coordination of time
overcurrent relays in this case can best be performed numerically
using tables or spreadsheets. In addition, spreadsheet techniques
may be used for setting other relays, such as feeder protection
relays. An example of the coordination of an industrial power
system with multiple generators connected through a synchro-
nizing bus, which a h has two utility ties, will be examined to
illustrate these points.
Index Tern-Industrial power distribution, protective device
coordination, protective relaying.

I. INTRODUCTION

T HE coordination of time overcurrent relays is the process

of determining settings for the relays that will provide
an orderly shutdown in case of a fault. Protective device
coordination is described in [1]-[3).
When multiple sources of power come into a bus (see
Figs. 1 and 2), the relays connected to each source will detect
a different amount of current for a feeder fault [6], [7]. A
current summation relay (“partial differential” scheme) on
all incoming sources will detect a still different amount of Fig. I. Relaying one-line Sagam ior bus.
current. After the motor contnbution has died out. the feeder
breaker relays will detect the 5ame current as the bus current
summation relays. There are two methods tha~haw been t ~ x d!o coordinate
The coordination problem is to determine the sequence of this type of system. The tint is to h u J umexurrent curve
relay operations for each possible fault location and to provide based upon per-unit fault current. For exampie. in Fig. 1. if
sufficient coordination margins without excessive time delay.
Relay 67 detects 4675 A. while Relay 5 1 V detects I86 1 A
A conventional time-current curve (Fig. 3). based upon the
for the same fault. the Relay 51 V curve is stufted to the right,
fault current detected by the feeder relay, will show conflicts
so that its 1861 A lines up with Relay 67‘5 -1675 A. This
in coordination where none exist. This is because the incoming
may easily be done when drawing curves by hand. The curves
line and tie-breaker and generator relays will each detect a
all stop at the maximum fault current. and the coordination
smaller amount of current than the feeder relay. Each of the
source relays will also detect a different current from the other margins are read along that vertical line on the graph. The
sources. computer-aided time-current i w e ploning programs which
are available generally have a current jcaie iabeled in amperes.
To plot a per-unit fault current curve. the pickup p i n t for each
Paper ICPSD 9-8. approved by the Power Systems Protection Committee
of the IEEE Industry Applications Society for presentation at the 19%
relay is adjusted to make the desired shifrs. and the current axis
Industry Applications Society Annual Meeting. San Diego, CA. October 6-10. is only accurate for one relay.
Manuscript released for publication March 14, 1997. The second method is to create tables that list the primary
The author is with the Power Systems Energy Consulting Department,
General Electric Company, Schenectady, NY 12345 USA. fault current detected by each relay. tbe relay current, the
Publisher Item Identifier S 0093-9994(97)054674. multiple of pickup. and the time delay of the relay for that
0093-9994/97310.00 0 1997 E E E
S w PROECTJVE DEVICE COORDINATION IN AN INDUSTRIAL POWER SYSTEM

115 W LmuTy SYSTEM

i
0 Bus1 Q 0 Bus2 BUS3 0 BUS4

I
13.1 kV PUNT SYSTEM
j

SYNCHRONIZING BUS

Fig 2 On?-line diagram ubed for example system in Section I1

current. The coordination margins may then be determined by 50/5 1T-the transformer primary time and instantaneous
subtraction. overcurrent protection;
5 1N-the residually connected ground fault transformer
primary time overcurrent protection (it does not enter into
the coordination problem discussed here);
11. EXAMPLE SYSTEM 5lGT-the transformer secondary ground fault backup
An example medium-voltage industnal power system is time overcurrent relay;
shown in the overall one-line diagram of Fig. 2. The over- 67T-a directional time and instantaneous overcurrent
current relaying is shown in Fig. 1. Each bus is supplied by a relay which operates for fault current flowing into the
generator. and the generators are different sizes. Each bus also transformer secondary cable:
is supplied by a transformer, through a set of long cables. The 67TN-a residually connected ground fault directional
four buses are tied together through current-limiting reactors time overcurrent relay which operates for fault current
to a synchronizing bus. Because of this arrangement, all flowing into the transformer secondary cable.
Coordination must be based upon the largest feeder on any To evaluate the coordination, the following fault scenarios
of the buses. The complement of relays to be considered here must be considered:
are the following: a feeder fault on the largest feeder with the highest relay
50/51F-the feeder breaker time and instantaneous over- settings;
current relay; a fault on the synchronizing bus:
5 0 G S t h e feeder breaker instantaneous ground sensor a fault on the transformer secondary, separated from the
relay: main bus by the secondary cable;
51B-the bus summation time overcurrent relay, con- a fault on the transformer primary.
nected to the penerator, transformer. and synchronizing The fault currents are calculated using 30-cycle multipli-
bus inputs; ers, because time-delay relays are being considered. Where
51BN-the residually connected ground fault bus sum- instantaneous relay settings are to be determined. the first-
mation time overcurrent relay, connected to the generator, cycle asymmetrical currents are calculated. The short-circuit
transformer, and synchronizing bus inputs; calculation program used must be able to determine fault
5 1V-the generator backup time overcurrent with voltage currents at remote bus locations. The program used here
restraint relay: produces output in one-line diagram format. as well as text
5 1G-the generator ground fault backup time overcurrent output, which aids in visualizing the results.
=lay:
'-
;* 6 7 4 dmxtional time and instantaneous overcurrent re- III. FAULTOK Bus 1-SOURCES ON Bus 1
lay lookmg into the synchronizing bus;
-67N-a residually connected ground fault directional time Relay coordination is determined using tables. The simplest
overcurrent relay which operates for fault current flowing case is for a fault on the Bus 1 feeder while it is k i n g
into the synchronizing bus; fed from the transformer only. This is shown in Table 1.
00001 OOotOOW 008 OW OW DOT 09 Ot 02 8 9 t 2 8 S o1
020 ’
OtD ,
090’
080 ‘
oot ’
002.
i..
009 ’
008.
00 ’1
00 ‘Z
0 0 ‘P
00 ‘9
00 ‘ 8
0 .OI
0 ‘W
0 ’Ot
0 ‘09
0 ‘08
’ 001
‘002
‘ow
’009
‘OOB
0001
 CURRENT I N AMPERES
...............
-wI
- 1s IR
" I lLoA PICKIP
IWQlElgD X 1.67
. ... .... ..
. .. . .. ..
....... ...."i & L..." ............... L .
i .
. .
. : .
.
.: .. .
. . . .
10 0 . *
.... .....
~ .!.. ......... .L
80 . 0 ....I .....i....:. ............i...............L.. .
:
.
_
;... .........r' :...
...a ...... .....,............ .+ .............. i
4

6 0 , 0 ,._T',...; ................... a.............. .....&..A

. . . .

40. 0 i.. ..... i............................................. i ......

. .
. . I
. .
. .
. . . . .
. .. .
. . . .. . . . : >
. ..
.
. . WrALUslEn

. . .........* . ............... :................... ....*........

. . . .
. . . ... . . .
... . .. . ...
.. . .. .. ..
.. ... :.. :
"
. .*-*... . ' d. " i..
... .. ..
b.."

..........,.... ................-..
..

. 600
. .. . .. ..
. ,. ... ..
....._._...-
.. :.. j j
... .... .... ....
- i
... ..
. .. .
. . ..
---.:-..:
.. .. ..
.
.. .. ...
.
;
i
.:
.. .... . .._....*
...-+
.. .. . .. .
. . ....."." i ............... ;
.. .. .. ..._..__...
..i.. .. .. ..:. :..
_j
. .. ..
i........ + i..
.. .. :
-
i . .

400 :...&.;.-~..-" .-.. . ,. . ..

"-...-^-".FT...C^.._.& ___.__ -. ....-..-........... 1 :____.* ._..,...., " j ....... .......,.._...- ....i..

.. . .. . .. . . . . .. ..
. . . . . .. .. . .... ... . . .
. .
: j i j j. j. j j .. .. . . . ... .... .... .... . . .
. . .. .. . . . .. .... ..
. . .. . . .
.. .. .. . . .. ..
-..i 1 . . . ... .. .- :--..-:.-...I i ^: ".+....c i-.............;.... ...... .....;...
I. . i .i .j . .
! ... ... .... ... .:. .:. !. :. j

.. ... ... . . . . . .
. .
. .i
i
.i ;
.
..2,-."--.4.-LL-
.. .. . .. . ....
* . -....
L. .
+_.."I__..& i ...... "I..... ..._.....
. i . .
.+.-.--1.-z.+
. .. .. ..
_..._ .: . .
: : !
+-.
.__..._.._.*.I_
... .. .:.-...:.-
. ...
.. .. .
.. .. : ..........j ... .... . !
.. . ._I .
040 ..I...i...i ;............... 4......... ,......l...;.."
... . ..
..?.
. .. .
: ;
.. ..
: :
: . . . . . . . . .
. . .. .. .. .. ... ....
.. .. ...
i . . ..--.:
. ..-.......-._._..
.... .... .... ....
+.._A

.. .... .. .. ...
.. ... . ... ...
."
..
68
.. . ... ....
i .....-. "..A
20
..... ..... .... ...
--..A,..;

40 60 100 200 400 800

....--_-.
-._
-.: : :
....:- >
i

20004000
i...-..A.
r.

10 OD
I loo IEEE TRANSACTIONS ON INDUSTRY APPUCATIONS. VOL. 33. N O 4. JULYlAUGUST 1997

R&P Fault Mulbpk of 1 TimctoClorc COordMnlpn

M Y Rclrup Dlal scumds seconds
s1v 960 4356 4.5 10.0 135 0.50
0% Vdtr
519 1200 13S6 3.6 30 I 30 0.45
SW51F FDR 400 4356 10.9 35 0 85
INST 4640 I

TABLE IV
PHASE OVERCURRENT COORDINATION ON BUS 4 FOR \ FWLT ON BUS I

A simplified assumption, which may result in slightly longer v. COORDINATION FOR A FAULT ON THE SYNCHRONIZING BUS
operating times than actually occur, is that the relay will detect The coordination should be checked for a fault on the
a fault current based upon the generator transient reactance synchronizing bus. This should be done separately for each
as determined during the 30-cycle short-circuit calculation bus. or set of identical buses, which feed the synchronizing
procedure. It would be helpful if computerized time-current bus. The synchronizing bus is protected by directional time
curve programs would calculate the true operating time of the overcurrent relays. These relays will operate for fault current
overcurrent relay along the generator decrement curve. The flowing into the synchronizing bus. They will mp for faults on
minimum relay operating time will occur for a fault near the the synchronizing bus, its reactors. or on a remote bus. These
bus. therefore, the relay curve for 0% voltage restraint should relays will not trip for a fault on the bus to which they are
be used. In this case, the fault current is less than the feeder connected. (In some schemes. synchronizing bus directional
instantaneous trip setting, and the coordination is done on relays may be used as differential relays by wiring their
that basis. The generator asymmetrical first-cycle fault current tnp contacts in qenes. This has the disadvantage of reduced
should be enough to cause a feeder breaker instantaneous reliability. because all ;days must function correctly for ;I
overcurrent relay to trip. The backup coordination deals with trip to occur., The coordination for this condition is shown
situations where this does not occur. The coordination 15 in Table V with .he remote fault. the 5 l B relay does not
simplified if the different buses are identical. but each bource jete:: !hi\ :3u;t. ind the coordination margins are long.
on each bus must be checked for any differences that prevent
the relay settings from being copied directly.
\ I. F A L L 3 OK THE TR.\NSFORVER SECONDARY CXALES
'The four 516 bus sumrmng relays provide overcurrent
Iv. COORDINATION FOR A REMOTE BUS FAULT protection for the 2200-ft 13.8-kV secondary cable. In ac-
The fault current for a fault on a remote bus will pass cordance with the National Electrical Code (NEC) [ S ] , trans-
through the synchronizing bus. Directional time overcurrent former primary protection is not considered to be adequate
relays (device 67) found on the source bus will back up the protection for secondary cables over 25 ft in length. The
bus summation relay (device 51B) on the remote bus. The 51B nondirectional overcurrent relays on the main 13.8-kV
worst case for coordination will occur at the maximum fault breakers provides overcurrent protection for the cable for
current, rather than at the instantaneous trip level, so this will current flowing through the incoming line. In accordance with
be calculated similarly to Table 11. Because this is a remote the NEC, medium-voltage circuit breaker protection can be set
fault, the currents seen by the source overcurrent protection at up to six times cable ampacity. A setting of 1200 A (less
may be below the pickup value. In addition, the source bus than two times ampacity) provides pickup coordination with
summation relay (device 51B) is not called into play because downstream relays. The time dial setting allows an adequate
this is a through fault. The sum of the fault currents entering coordination margin with the largest downstream feeder and
through the generator and transformer equals the fault current with the upstream 50/51 incoming line relay.
leaving through the synchronizing bus. Examination of Table For fault current supplying the cable from within the plant,
IV shows that the coordination margins for this condition are either through the generators or the synchronizing bus, a
well above the 0.3% minimum requirement. directional overcurrent relay, device 67T,is used. This relay is
S w : PROTECTIVE DEVICE COORDINATION IN Ah! INDUSTRIAL POWER SYSTEM 1 I01 -

TABLE V
PHASE OVERCURRENTCOORDINATION
ON Bus 1 FOR A FAULT ON Bus SYNC
I Pickup I Fault I Multiple of I Time I T i m to Close I coord.Margin
Relay I A m P m S l ~ IPjcLup I Did I seconds I secondr
WSIT I 160 I 338 I 2.1 I 3.5 I 2.25 1 1.5

BUS 1
05€ Volts
67 SYNC 960 467s 4.9 2.0 0.74
BUS 1 INST. 5400

Rckup Fault Multlple of Time ‘IimetoClose Coord Margn

Relay Ampms Ampem Rckup hd seconds &on&
67 T 1 200 7.548 63 10 0.35
INST
I2800
SlV 96c) 2334 j 3 100 IS 115
BUS 1 I

set to carry the maximum full-load current of one transformer. the ground fault relays is checked using the same techniques
for the case in which the generator is supplying the 115- as discussed above for phase faults.
kV system. The instantaneous setting is above the maximum
asymmetrical fault current detected for a 115-kV system fault. VIIl. FEEDER BREAKER RELAI
The coordination must be examined for a fault on the PICKUPS USING SPREADSHEETS
transformer secondary cables when the fault is fed by the local The feeder breaker relay settings for this plant were de-
bus, which means the generator. This is shown in Table VI. termined based upon the protection of the feeder cables and
The coordination must also be examined for a fault fed b) a unfused distribution transformers on the feeders. The time and
remote bus. This is shown in Table VII. instantaneous pickup points were determined using a spread-
sheet. The time delay settings were found using time-current
VII. GROUNDOVERCURRENT CWRDINAnOlV curve coordination with substation fuses. breakers. transformer
Coordination for ground fault protection will use the same inrush points. and transformer ANSI curves (not shown).
techmques, except that the ground fault currents are limited by Tables VIIl ( a H d ) illustrate typical column headings and the
the grounding resistors at each source [7]. The ground fault settings for two feeders. The time overcurrent pickup minimum
current may then be determined either manually. using the setting should carry the ampacity of the cable. The maximum
current from each resistor, or with a computer. calculating the setting should be below the maximum allowed by the NEC
short-circuit currents. The results in most cases should not for cable and transformer protection. These minimum and
differ significantly. Each feeder has an instantaneous ground maximum pickup settings are calculated automatically. Two
Sensar relay. The first backup is a S1BN bus summation relay. cable sizes are evaluated. the smallest size present on the
W k h is set low in order to be sensitive to a wide range of feeder and the size of the firs! section of feeder that leaves
a t currents. The ground backup relays on the transformer the breaker. The cable ampacin is denved from a lookup
md generator. devices 51G and 51GT. are set the same and table. In this case. all cables type\ were identical. and the
cOOrdinate at pickup with the 5 1BN relay. The coordination of lookup could be from one NEC table copied into another
 -
IEEE TRANSACllONS ON INDUSTRY APPUCAnoNS. V O L 33. NO. 4. JLIY/AUGU!3T 1997

TABLE vnr
FEEDER RELAY PICKUP T ABLE (a). tb, TIMEOVERCURREST P!CKUP SPRWOsHEET I c )
INSTANTANEOUS OVERCURREKT PICKUP SPRWUSHEET id) CAB= PROTECTION SPREADSHEET

i a)

Max. Ampcrcity A0 Roc.

Found
Bus Feeder CT Tap Tap Tap Tap Amp ChWW
PU

1 A 80 11 5 8 5 400 IChanm
1 B 6 0 1 17 5 1 4 1 4 32010k

w Max. Max. Inlt.

CT T m . Trans. Sym. Aeym. Sa
Bum Feed.r R.li0 kVA %z I& I= Amps
1 A 80 3750 5.48 2873 5056 53
1 0 m 1500 5.63 I 1115 1928 25

TABLE IX
PICKUP CALCULATION FOR A GENERATOR 51-V RELAY ( a ) PICKLP WITH FULL RESTRAINT SHOULD EXCEED 1508 OF
GENERATOR F ULL L OAD CURRENT (b) RCKUP WITH ZERO RESTRAINT SHOULD B E B ELOW MINIMUM FALXTCURRENT

51V 51v
Full Load Pickup Pickup
Gemrauw Amps(FLA) 1.5xFLA CTRauo 100% V iOO% V commnts
c 573 7u5 1 M 6 ! 31511 1 ilk

Full 5iV I ilV '

Load Min. Pickup RcKup I

Genemar Amps(FLA) Xd FauItArnps CTRaho O%V O%V 1 Comments

J 523 1.3 I 402 160 15 2 4 0 1 ok

section of the spreadsheet (not shown). The recommended The maximum trip time. in order to protect the cable Z 2 t
tap setting was determined manually, although it could be heating limit, can also be calculated [ 2 ] .
programmed, if desired. The final column flagged a difference Where fuses are present, the table can be augmented to
between "as found" and "recommended" settings. The second assure that the relay pickup current is above the largest
part of the table covers instantaneous pickup. In this case, fuse clearing time curve at loo0 s for "E"-type fuses or
the instantaneous unit is to be set above the maximum first above the continuous current rating for "R'-type fuses. Where
cycle asymmetrical fault current for any transformer secondary medium-voltage motors are present. the instantaneous pickup
bus on the feeder. The symmetrical current is first calculated. should be above the maximum asymmetrical inrush current.
The choice of a multiplier in the worst case would be 1.1 x This may be determined as 1.6 x 1.1 times locked rotor
1.6. allowing for maximum asymmetry plus an overvoltage current and included in the spreadsheet in the same manner as
tolerance factor. In many cases, the asymmetry is less due to transformer secondary fault current. The coordination should
a lower X/R ratio than that which produces the 1.6 multiplier. also be checked with curves in these cases.
SVIHERLAND. PROTEOWE DEVICE COORDINATION IN AE; INDUslTuAL POWER SYSTEM 1103

Ix. GENERATOR 51v RELAY PICKUP REFERENCES

The generator backup phase overcurrent protection is pro- [ 11 C. R. St. Pierre and T. E. Wolney. “Standardization of benchmarks for
vided by voltage-restrained time overcurrent relays (device protective device timecurrent curves,” IEEE Trans. Ind. Applicar.. vol.
IA-22. p ~ 623-633.
. JUly/AUg. 1986.
51V). The pickup of these relays is set such that the relays 121 IEEE Recommended Pracrice for Electric Power Disrriburronfor Indus-
will pick up with full restraint above the maximum generator rrial P h r s . IEEE Std. 141-1993.
overload current, assumed to be 150% of full-load current. [3] IEEE Recommended Pracrice for Pmrection and Coordinarion of Indus-
rrial and Commercial Power Svstems. IEEE Std. 242-1986.
With zero restraint, the relays should pick up at a maximum [4] Industrial Power Svsrems Dora Book, General Electric Co.. Schenectady.
of U3 of the fault current supplied through the generator NY. 1%8. sect. 54. pp. 61-65.
synchronous reactance with no field forcing (“stuck regulator“ 151 Narional Elecrrical Code. ANSVNFPA 70, 1993.
161 R. H. Simpson, “Protective relaying for multi-source generator buses.”
condition). These settings are illustrated in Tables IX-A and IEEE Trans. Ind. Appbcar.. vol. 26. pp. 330-341. Mar./Apr. 1990.
B. The 51V relay time curves are set to coordinate with the 171 F. S. Prabhakara. R. L. Smith Jr.. and R. P. Swatford. Indusrrial
and Commercial Power Systems Handbook. New York: McGraw-Hill.
largest feeder breaker and with the directional relays protecting 1995. pp. 8.54-8.59.
the synchronizing bus and the incoming line.

Peter E. Sutherland lS’77-M’79, received the

X . C O NC L US IO N A.S. degree in electncal engineering technolog! and
the B.S. degree in electncal engineenng fror: the
The methods for coordinating a multiple source system with University of Mame. Orono. in 1979 and lYS3
time overcurrenl rela! 5 were presented 4 n example using respectivel!. and the M.Sc.E degrre in elei‘tncal
the spreadsheet or Labie caicuiatror: methd wa3 presented. engineering from the Llniverrlt! ot Fveu Biunsuich.
Fredericton. N.B.. Canada. in 1986 Hc 1, cur-
When coordinating such a system. man! factors mus: be rentl? working toward the Ph.D degree i n electnc
considered for setting each overcurrent rela?. Rather than power engineering at Rensselaer Polytechnic Insit-
attempting to remember each factor to be considered for each tute. Troy. N\’.
He was a Test Engineer and a Design Engineer
relay. it is preferable to have a systematic method where all with Accutest Corp.. Chelmsford. MA. a manufacturer of automatic test
factors must be considered. The use of tables simplifies this equipment for the semiconductor industry. He was also a Planning Engineer
requirement. for an electnc utility company for a brief period. In 1987. he joined
Genera! Electric Company. where he has been an Engineer in the lndusuial
Power Systems Engineering Operation in Schenectady. NY. an Instructor in
the Training and Development Center. and a Power Systems E n p e e r in
ACKNOWLEDGMENT Albany. hi.He is currently a Senior Engineer in the Power Systems Energy
The author would like to thank B. W. Cable of GE Power Consulting Department. Schenectady. NY. where he performs a variety of
power system studies for industrial clients.
Systems Energy Consulting. Schenectady. Nk’. who assisted MI Sutherland is a member of Eta Kappa Nu and Tau Beta Pi He is a
in the original study upon which this paper is based. Registered Professional Engineer in the States of Maine and New Yo&.

View publication stats