The 19th Electrical Power Distribution Conference (EPDC2014), 6-7 May, 2014, Niroo Research Institute

Optimal Coordination of Distance and

Over-current Relays Using ICA

M. Rahmatian1,2 M. H. Pourarab1 M. Gholami2 G. B. Gharehpetian2

1 Mashhad Electric Energy Distribution Co., 2 Department of Electrical Engineering
Mashhad, [ran Amirkabir University of Technology
Tehran, [ran
[email protected] [email protected]
[email protected] [email protected]

Abstract- Coordination of different relays in transmission optimization algorithm is discussed. The method is tested using
systems is an important issue in relays application. In this paper, a 6-bus network and the obtained results are shown in section
the over-current and distance relays are coordinated through the IV.
Imperialist Competitive Algorithm (ICA) considering time
setting multiplier (TSM) and pickup current of the relays. The II. THE PROPOSED METHOD
proposed method is investigated on a 6-bus network. It is shown
The objective function presented in this paper is used to
that ICA has better results in comparison with Genetic
coordinate the distance and over-current relays. The ICA is
Algorithm (GA).
applied to solve the problem as a heuristic algorithm. This can
Index Terms-Distance Relay, Over-Current Relay,
considered as a novelty of the paper. The obtained results can
Coordination, Imperialist Competitive Algorithm (ICA).
be compared with the results of other schemes. [n this paper,
I. INTRODUCTION similar to [1], five critical points are selected for coordination
of over-current relays together and distance relays with over­
Power systems are composed of different parts, including
current relays.
generation, transmission and distribution. These parts should
be protected against any fault for a continuous and safe energy [n this paper, TSM and pickup current of the relays is
supply. For instance, the transmission system should be considered for coordination, and this can considered as second
protected by different relays mostly distance and over-current novelty of this paper. The operating time of over-current relays
relays. These relays protect the transmission system against the with standard inverse characteristic can be obtained as follows:
over current.
0.14
t=TSM(
One of the most considerable aspects of these relays M002 -I ) (1)
application is the coordination between them to avoid the
where M is the ratio of short-circuit to pickup current of the
unnecessary trips. Therefore, this issue has been under study
relay. The TSM parameter is considered as a continuous
and different methods are proposed such as dual simplex [1]
parameter in the range of 0.5 to 2 while the pickup current is
and simplex method [2]. Furthermore, some methods have
considered as a continuous current in the range of 1.2 to 1.3.
been proposed for distance relays coordination e.g., linear
programming [3]. [n [4] linear programming has been used to The objective function used in this method is as follows:
coordination of directional over-current relays. In [4] defmite
time backup relays have been considered. [n [5] hybrid PSO OF = OFuc_oc + OFuc-dis + OFdis-uc (2)
has been used to coordination of over-current relays. In [6] The first term of (1) belongs to the coordination between
hybrid particle swarm optimization (PSO) has been used to the over-current relays. This term is composed of:
coordination of over-current and distance relays.
.v FJ 2
[n [7], a new GA-based method for coordination of distance OFoc_oc aL)o + PI kL(L�tmboCifil -I�tmbocljodl)
l
=

n=[ =[
and over-current relays is presented. In this study, the
characteristics of the relays and TSM are chosen as the P 2
p,I(�tmboChl -lillmbochll)
optimization variables to coordinate the over-current and +
k] =1
distance relays.
P 2
In this paper a new method is proposed to coordinate of +P2I
k] (�tmh
=1
+ �tmh
()'"iFll l ()Wlll)
distance and over-current relays using the [CA. [n this method,
TSM and pickup current are selected. A standard inverse P. 2
characteristic is assumed for the relay and the objective +PJI(�
kl�' tmhl)qF21 +1�tmb()W211) (3)
function is similar to the objective function presented in [7].
where N is total times of the over-current relay operation,
This paper is organized as follows; the proposed method is PI is the number of the main and backup over-current pair and
presented in section II. In section III, the ICA as the applied

978-1-4799-563 6-4/[4/$31.00 ©20 [4 IEEE 106

�1, �2 and �3 are the weighting factors chosen based on their where P3 is the number of over-current and distance relays
importance. as the main and backup relays, �6 is the weighting factor, tz2 is
the operating time for the second zone of distance relays, CIT3
The first term of (3) is the sum of the operating times of
is the desired time interval between the operation of over­
over-current relays originated from close-in faults. The other
current and distance relays, tm(F5) is the over-current relay
terms are added to minimize the time interval for fault
operating time for the fault occurring at F5 and tz is the
discrimination by the main over-current relays and backup
standard time delay of the distance relays for the second zone.
M
ones, i.e., mbOCIF,1 is the time difference for fault III. IMPERIALISM COMPETITIVE ALGORITHM
discrimination between the main over-current and backup relay ICA was introduced by Atashpaz and Lucas [8]. This
for the fault at Fi. This parameter is determined as follows: method is a new socio-politically motivated global search
strategy which is able to deal with different optimization tasks.
(4)
Like all the evolutionary algorithms, this algorithm starts
where CTIl is the desired time interval between the over-
with an initial population of P countries. These countries are
current relays operation as the main and backup relays.
generated randomly within the feasible space. Each country
The second and third tenns of (1) belong to the can be shown as follows:
coordination of the over-current and distance relays. The
countryi = [pI, p2, ... , pNvar] (10)
difference between these tenns is in the assumption of the
backup relay, i.e., the backup relay in the second tenn is over­ where Nvar is dimension of the optimization problem. The
current relay while in the third term the backup relay is the best countries in the initial population are selected as the
distance relay. imperialists based on their cost function. Other counties are
known as the colonies of these imperialists. These colonies are
The second term is stated as follows:
distributed among the imperialists based on imperialist's power

L (L'l.tmboCDlSlr,1 - 1L'l.tmboCDlshl l) 2
P,
o generate the initial empires. Therefore, the normalized cost of
OFoc_di., = PI an imperialist is defined as follows:
kz=l

f (Mmh()cTJJSIF41 - IMmb()cTJJW41 1) 2
p
Cn= cn - max (ci) (11)
+ PI
kl=l where cn, Cn are the cost and the nonnalized cost of nth

L (L'l.tmboCDlshl l L'l.tmboCDlshl l) 2
1, imperialist. Consequently, the normalized power of each
+ P. + imperialist can be defined:
k2=[

L (Mmh()CTJJSIF41 l L'l.tmh(X"TJJSIF41 1) 2
P,
C
+ /3, +
Po=I � 1
k2�1
(5)
where P2 is the number of distance and over-current relays ! C,
j=[
as the main and backup relays, respectively and �1, �4 and �5 (12)
are the weighting factors.
where Nimp is the number of the empire colonies.
S·Iml·Iar to the fiIrst tenn, Lltmbocmslrl ·
' ·IS the tIme d·f'"
I lerence After spreading the colonies among the imperialists, they
in fault discrimination between the distance relay as the main begin closing to their consequence empires. The total power of
relay and the over-current relay as the backup relay for the each empire can be calculated as follows:
fault at Fi and is follows: Powern = cost(imperialistsn) + � mean {(cost(colonies of
empiresn))} (13 )
(6)
where CTI2 is the desired time interval between the
operation of distance and over-current relays as the main and where powem is the total power (cost) of the nth empire
backup relays, respectively. and � is a positive number considered less than 1. It states that
the power of each imperialist country plus a percentage of
The third tenn is as follows:
power of its colonies forms the total power of the empire. In
p] this paper, � is set to 0.1 considering [8].
0F;!is-oc = j3e; L (LltmbDISOC)
k3=1
(7)
where, During the competition, each empire that is not able to

LltmbDlSOC = tz2 - CTI3 (8)

increase its power will be removed from the competition.
Consequently, the power of strong empires will increase and
and, the weak ones will weaken more. Finally, the weak empires
will lose their power and collapse. At the end of this process,
(9) one empire will exist which is the answer of optimization
problem. This algorithm is depicted in Fig.l.

107
 The movement mechanism of the colonies toward their As shown in Table II, there are 20 peers of the main and
consequence empires and other detailed descriptions about this backup relays, i.e., PI=P2=P3=20. The results of the presented
algorithm can be found in [8]. algorithm are shown in Table III. The amount of TSM and
pickup current are shown for each relay in this paper.

In Table IV, the results of the proposed method are

compared with those obtained in [7] through GA. The pickup
current is assumed 1.2 times of the load current. As it is shown,
the obtained amounts for the objective function through ICA
are better in comparison with those obtained through GA.

TABLE II- THE SHORT CIRCUIT CURRENTS OF FIVE CRITICAL FAULTS

t:!:l t:!:l t:!:l t:!:l
Fig. 1. The imperialist competition algorithm � Il> � Il> Il> Il> �
(") (") (") (")
Il> Il> Il>
:>;" :>;" :>;" :>;"
t:!:l "Tj 5" "Tjl" "Tj 5" "Tjl" "Tjl" "Tjl" "Tj 5"
� Il>
Il>
::>;j 1l>'O Il>
::>;j 1l>'O 1l>'O 1l>'O Il>
::>;j
(") S('[) S::>;j S('[) S::>;j S::>;j S::>;j S('[)
Il>
:>;" M Il> - M ('[) M Il>- M ('[) M ('[) M ('[) M Il>-
5' I" Il> Il> - Il> Il> - Il> - Il> - Il>
IV. SIMULATION RESULTS '0 M '< M il> M '< M il> M il> M il> M '<
::>;j "TjC/) "Tj.. '< "TjC/) "Tj'< "Tj'< "Tj'< "TjC/)
('[) ::>;j ...... n .... C/) N n NC/) wC/) >l'-C/) Ul n
The proposed method is simulated on a 6-bus network SJ ('[) ,-.,n n ,-.,n n n n ,-.,n
'< SJ >-1" ..,
S;:n >-1" ..,
S;:n S;:n S;:n >-1" ..,
shown in Fig.2. '< '-.-/ .., '-.-/1" '-.-/ .., '-.-/1" '-.-/1" '-.-/1" '-.-/ ..,
('[) .., ('[) .., .., .., ('[)
.., .., .., ..,
::l ('[) ::l ('[) ('[) ('[) ::l
M ::l M ::l ::l ::l M
M M M M

2 1 5428 828 3544 386 828 455 3994

14 1 4184 816 1607 816 816 488 1584
3 2 3505 3505 1789 1789 3505 2054 2361
4 3 1769 1769 1117 1117 1769 1244 1521
5 4 1103 1103 354 354 1103 519 733
6 5 4936 340 2719 19 340 75 3146
7 5 4184 337 1607 337 337 202 2991
1 6 2682 2682 853 853 2682 1216 1685
'J T
'5 2 7 5428 1571 3544 734 1571 864 4017
(] 8 7 4933 1563 2529 59 1563 226 3503
13 8 2492 2492 852 852 2492 1180 1654
8 9 4933 340 2529 13 340 49 2958
14 9 4184 337 1607 337 337 202 1584
Fig. 2. The schematic of the 6-bus network
9 10 1174 1174 355 355 1174 536 867
10 11 2589 2589 1195 1195 2589 1445 1703
11 12 3655 3655 2615 2615 3655 2786 3052
The network under study has 7 lines. Consequently, 14 7 13 4184 816 1607 816 816 488 3169
distance relays and 14 over-current relays are installed in the 12 13 5431 828 3681 421 828 485 4114
network. The load current of the over-current relays are listed 6 14 4936 1565 2719 87 1565 345 3592
in Table I while Table II shows the short-circuit current for five 12 14 5431 1573 3681 799 1573 921 4142
critical faults of the network [7].
TABLE 1- THE LOAD CURRENTS OF OVER CURRENT RELA YS
Load Load
Relay No. Relay No.
Current Current
1 104 8 109
2 166 9 118
3 125 10 110
4 180 11 135
5 129 12 122
6 114 13 125
7 141 14 166

108
 TABLE III- THE RESULTS OF THE PROPOSED METHOD [8] E. Atashpaz-Gargari, C. Lucas, "Imperialist Competitive Algorithm: An
ICA Method Algorithm for Optimization Inspires by Imperialistic Competition,"
Relay IEEE Congress on Evolutionary Computation,CEC 2007, Singapore, pp.
Pickup
No. TSM 466 1-4667, 25-28 sept. 2007.
Current
1 0.185 1.2
2 0.178 1.2
3 0.150 1.2
4 0.102 1.2
5 0.086 1.2
6 0.301 1.25
7 0.113 1.2
8 0.266 1.2
9 0.079 1.2
10 0.113 1.2
11 0.130 1.2
12 0.205 1.21
13 0.143 1.2
14 0.136 1.3

TABLE IV- COMPARISON BETWEEN ICA AND GA

ICA GA
Objective 4346.1 4515.2
Function

V. CONCLUSION

A new method for coordination of distance and over­

current relays has been proposed in this paper. The parameters
considered for the optimization are TSM and pickup current of
the relays. The validity of the proposed method is investigated
through a 6-bus network. The obtained results are then
compared with the results of OA. It is proved that the ICA has
better results.

REFERENCES

[I] A. J. Urdaneta. H. Resterpo. J. Sanchez. and J. Fajardo. "Coordination of

directional OIC relays timing using linear programming,"' IEEE Trans.
Power Del., vol. 1 1, no. I, pp. 122-129, Jan. 1996.
[2] H. Askarian Abyaneh and R. Keyhani, "Optimal co-ordination of OIC
relays in power system by dual simplex method," presented at the
AUPEC Conf., Perth, Australia, 1995.
[3] L. G. Perez and A. J. Urdaneta, "Optimal coordination of OIC relays
considering definite time backup relays," IEEE Trans. Power Del., vol.
14, no. 4, pp. 1276-1284, Oct. 1999.
[4] L. G. Perez and A. 1. Urdaneta, "Optimal coordination of OIC relays
considering definite time backup relays," IEEE Trans. Power Del., vol.
14, no. 4, pp. 1276-1284, Oct. 1999.
[5] M. Bashir, M. Taghizadeh, 1. Sadeh, H. R�jabi Mashhadi, "A New
Hybrid Particle Swarm Optimization for Optimal Coordination of Over
Current Relay" International Conference on Power System Technology,
20 10
[6] J. Sadeh, V. Amintojjar, M. Bashir, "Coordination of Overcurrent and
Distance Relays Using Hybrid Particle Swarm Optimization", The
International Conference on Advanced Power System Automation and
Protection, 20II
[7] R. Mohammadi Chabanloo, H. Askarian Abyaneh, S. S. Hashemi
Kamangar, and F. Razavi, "Optimal Combined Overcurrent and
Distance Relays Coordination Incorporating Intelligent Overcurrent
Relays Characteristic Selection", IEEE Trans. Power Delivery, vol.
26,no. 3,.1ul. 20II

109