The Journal of Supercomputing (2021) 77:10721–10741

https://doi.org/10.1007/s11227-021-03705-8

Optimal coordination of over current relay using opposition

learning‑based gravitational search algorithm

Debasis Acharya1 · Dushmanta Kumar Das1

Accepted: 21 February 2021 / Published online: 10 March 2021

© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature
2021

Abstract
In this paper, the optimum coordination of over current relay in distribution system
has been performed. The time dial setting (TDS) of over current relay is optimized
by minimizing total operating time of relays. As the power system is a complex and
highly interconnected, the complexity may be occurred in calculating optimum relay
coordination time (TDS, phase setting etc.) by conventional or analytical methods.
Therefore, an opposition learning-based gravitational search algorithm (OLGSA) is
proposed for relay coordination optimization problem. The proposed optimization
method can search the optimum solution for relay coordination problem from both
the direction (initial populations and their opposite direction populations) in search
space to reach global solution as early as possible. Three different distribution power
systems are considered to verify the performance of the proposed algorithm for
optimum relay coordination problem. The simulation results for all three cases have
been compared with existing results. The outcome of the proposed method for relay
coordination problem shows that the total relay operating time is minimized (20–40
% improvement in relay total operating time for different cases) comparably than
available results for all three cases by maintaining coordinations among relays and
satisfying different constraints.

Keywords Time dial setting (TDS) · Plug setting (PS) · Coordination time interval
(CTI) · Opposition learning-based gravitational search algorithm (OLGSA)

* Dushmanta Kumar Das

[email protected]
Debasis Acharya
[email protected]
1
National Institute of Technology Nagaland, Dimapur, India

13
Vol.:(0123456789)
10722 D. Acharya, D. K. Das

1 Introduction

An uninterrupted power supply to the consumer is a challenging factor in power

sector. In restructure power system environment, the financial issues have a
great impact on continuous power supply. In such scenario, the protection sys-
tem in power supply is a vital part to maintain power supply smooth and keep
the financial revenue balance correct. Improper use and activity of power system
protection devices can lead an unbalance power supply and cause loss in finan-
cial revenue in power industries [14]. Over-current relays are extensively being
used in power system distribution network as primary protection devices. It is
also applied for backup protection in transmission and distribution system. Such
relays are best in selectivity, sensitivity and accurately [5].
Generally, the primary protections are set to operate fast when there is a fault
happened in the zone of primary relay network. It makes the circuit breaker (CB)
open and isolates the faulty part from the rest of the power system. Backup pro-
tective relays are set in such a way that it must operate only if the primary relay
does not work in its assigned time limit. The backup relay must follow the coor-
dination time interval (CTI) after initiating the fault in its zone of protection [13].
Therefore, when there is a fault in distribution network, the primary relay in fault
zone must operate fast. Otherwise, after a pre-defined time interval, the back-up
relay operate to open the CB to isolate the faulty zone from the rest of the power
system. Correct coordination of primary and back-up relay is required in power
sector to supply continuous power and also maintain financial revenue. The coor-
dination between primary and secondary relays should be such that fast and accu-
rate fault clearance can be achieved. For directional overcurrent relay coordina-
tion problem, the total operating time of primary relays must be minimum by
following coordination among other relays. The robust time dial setting is also
required to adopt all possible operating condition. Therefore, the optimization
algorithm should be such that it must find global optimal solution with fast and
accurate response from protective relays.
In conventional method, optimum-coordination of relays was done manually
by setting plug setting and time setting with expert knowledge or calculating ana-
lytically. Now-a-day, the power distribution network becomes so complex that the
conventional method of relay coordination is not a good choice. Therefore, opti-
mum coordination of relays is required to adapt [5]. In [25], graph theory and
optimization method have been discussed for the relay coordination purpose. In
[12], trial and error method was applied with a large number of iterations rate that
made the convergence rate slow. Some classical techniques such as curve fitting
technique [3], analytical-based technique [19] are proposed for relay coordina-
tion problem. Though, relay coordination problem is a nonlinear problem as time
dial setting (TDS) and plug setting (PS) required adjusting, with a fixed PS value,
it can be assumed as linear optimization problem [5]. In [24], linear program-
ming-based method for relay coordination problem is considering disjunctive-
inequalities. The linear approach to solve relay coordination problem suffers from
local optimal as TDS is verified keeping PS constant. Presently, the swarm-based

13
Optimal coordination of over current relay using opposition… 10723

metaheuristic algorithms are applied in different types of optimization problem

in different engineering field. There are many research articles performed in
relay coordination problem with the help of machine learning-based evolution-
ary algorithm such as genetic algorithm (GA) [2], particle swarm optimization
(PSO) [10], modified teaching learning optimization [9], ant colony [23], differ-
ential evolution [22] have been investigated for relay coordination problem. Some
other optimization technique like harmonic search algorithm (HAS) [16],chaotic
firefly algorithm (CFA) [7], hybrid grey wolf optimization (GWO) [12] have been
applied for relay optimization problem. In [3], hybrid GA-NLP method was pro-
posed for relay coordination problem. In [11], firefly algorithm for overcurrent
relays is used for over current relay coordination problem. In [14], water cycle
algorithm (WCA) have been tested for relay coordination problem.
Most of the available metaheuristic algorithms work for optimum relay coordina-
tion are based on randomly generated initial populations where the TDS of relay
is considered as variable keeping PSM constant [2, 14, 16]. The initial popula-
tions are randomly generated in one direction in the search space. As the position
of particles may be far away from global position, the convergence process will be
slow. The total operation time achieved in present literature can be made better with
proper relay-coordination. Therefore, the searching way of optimum solution can be
improved by taking the opposition direction solutions [26] of initial randomly gen-
erated populations and their updated solutions in each iterations. Thus, the better
global solutions can be obtained quickly and the total operating time of relay system
can be minimized than existing result maintaining all constraints. With such motiva-
tion, an opposition learning-based metaheuristic algorithm have been proposed for
optimum relay-coordination.
In this paper, the relay coordination problem is considered as optimization prob-
lem to find optimum TDS setting and minimum operating time. For this purpose,
opposition learning-based gravitational search algorithm (OLGSA) is proposed as
optimization algorithm. GSA is introduced in [17]. GSA provides fast convergence
with global solution [17]. The opposition learning concept [21] is merged with GSA
to enhance the efficiency of the proposed algorithm. To verify the effectiveness of
OLGSA in relay coordination problem, the performance is compared with some
existing results.

2 Related work

2.1 Over current relay coordination in power system

Consider a radial distribution system with two buses as shown in Fig. 1. If a fault is
initiated at point K near bus B, the relay R2 nearer the fault section must operate to
isolate this section from the rest part of the network. If the relay R2 does not operate
for any reason, the relay R1 will work as a back-up relay. In such case, the setting of
relay R1 must be higher than the relay R2. In general, the setting of relay R1 is addition
of the operating time of relay 1, circuit breaker (CB) operating time at a bus and the
overshoot time of relay R1. Consider a single fed ring main distribution system with

13
10724 D. Acharya, D. K. Das

Bus K Bus L

Relay 1
Relay 2
Fault F

Load Load

Fig. 1  Radial distribution system with two buses

eight relay. In this scheme, relay 1 and 8 are bidirectional relay and rests of relays are
unidirectional over current relay. The relays must work within their protective zone. In
this case, relay 2, 4, 6 and 8 will form a group and relays 1, 3, 5 and 7 will form another
group. The coordination between operating time of relay group 1 will be as follows:
TR8 > TR6 > TR4 > TR2 . Similarly, the coordination between operating time of relay
group 2 will be as follows: TR7 > TR5 > TR3 > TR1 .

2.2 Relay characteristics

There are different type of over current relay available. Generally, inverse definite mini-
mum time relay is used as over current relay because of its best characteristic [5]. The
operating time of such relay is defined as follows:
𝛽 × TDS 0.14 × TDS
Top = =
(PSM)𝛼 − 1 (PSM)0.02 − 1
, (1)

where Top is the operating time of relay, 𝛽 and 𝛼 are constant presented in Table 1,
PSM is plug setting multiplier. It is defined as follows:
If
PSM = , (2)
Ipr

where If is fault current and Ipr is relay peak-up current. For a constant PSM with
pre-define relay pick up current, (1) can be written as follows:
Top = ai × TDS, (3)
where ai = 0.14
(PSM)0.02 −1
.

Table 1  Values for 𝛽 and 𝛼 Relay type 𝛽 𝛼

of different over current relay
characteristics [18] IDMT 0.14 0.02
Very inverse 13.5 1
Extremely inverse 80 2

13
Optimal coordination of over current relay using opposition… 10725

2.3 Problem formulation for overcurrent relay coordination optimization

problem

In this paper, the coordination among over current relays of radial distribution
system is considered as optimization problem. The objective for the optimization
problem is to minimize the total operating time of all relays under the in a power
system network.

2.3.1 Objective function for overcurrent relay coordination optimization problem

For the optimum relay coordination problem, the minimization of total operating
time is considered as the objective function. It is defined as follows:
N
∑
min z = 𝜔j × Tik , (4)
i=0

where N is total number of relays, Tik is operating time at ith relay at fault k, wj is the
weight factor of network length. For distribution system, w is 1.

2.4 Constraints relay coordination

In this section, various constraints associated with relay coordination are pre-
sented as follows.

2.4.1 Operating time constraints

Relay must operate only when there is a fault occurred in its zone of protection.
Due to the variable load connected with a distribution system, there may be a
chance to happen over-current for switching on or off the load in the network.
Such over current vanishes within few cycles. Therefore, relay should not oper-
ate within such condition. The relay operating time should maintain a range of
operating time as follows: Tik min ≤ Tik ≤ Tik max where Tikmin and Tikmax are lower
and upper boundary of relay operating time. Typical values for Tmin = 0.1s and
Tmax = 1.1s [11].

2.4.2 Coordination time interval (CTI) constraints

The backup relay must operate only when primary relay does not work in its protec-
tive zone under fault condition. The operating time of backup relay must satisfy CTI
constraints defined as follows:
Tbk − Tpk ≥ CTI (5)

13
10726 D. Acharya, D. K. Das

where Tbk is backup relay operating time at fault k, Tpk is primary relay operating
time.

2.4.3 Time dial setting (TDS) constraints

The time dial setting of over-current relay should be bounded in a range such that
fast operation can be achieved with ignoring false operation due to switching exter-
nal load. It is expressed as follows:
TDSmin ≤ TDS ≤ TDSmax , (6)
where TDSmin is the minimum value of TDS, TDSmax is the maximum value of TDS.

Remark 1 The relay operating time depends on many factors such as proper selection
of primary and backup relay, the distance covered by relay system, proper setting of
relay in terms of relay speed of operation [22]. The distance covered by the relay
is adjusted using time setting called time setting multiplier. The relay operational
speed or how fast the relay operate affects on relay operational time. The speed of
relay moving part is associated with the force on moving part that causes due to flow
of current (equivalent fault current level) through the relay coil [5]. Thus, relay oper-
ating time is affected by the plug setting multiplier (PSM). Therefore, wrong setting
of PSM and TSM can effect on relay operating time. Thus, the relay operating time
get affected by setting of current, level of fault current, time setting and plug setting.
Some other factors are also the total operating time of relay like operational time of
primary relay, time interval in terms of coordination time of backup relay, circuit-
breaker operation time, overshoot and signal travels time etc. As, with the increas-
ing size of distribution network, the system becomes more complex, the coordina-
tion between relays becomes complicated. If the coordination is improper, the relay
operation will not be accurate. Poor performance of relay can lead malfunctioning of
circuit breaker (CB). Therefore, exact relay coordination with proper TSM, PSM set-
ting and minimum operating relay time is required in power network [24].

Remark 2 The purpose of relay system is to protect the power system from faulty
zone. It should operate quickly to avoid power outage in the system. The primary
relay system is required to operate fast in the protected zone which is also covered
by backup relay system. The backup relaying system must wait for a specified time
delay. As, the relay operating time depends on many factors like operational time of
primary relay, time interval in terms of coordination time of backup relay, circuit-
breaker operation time, overshoot and signal travels time, the total relay operational
time is large enough that can cause more delay operation on protective system and
can damage the equipment in electrical network. It can also reduce the efficiency of
the equipment. Therefore, the relay system must protect the power network as quick
as possible with maintaining different constraints like CTI, etc. The reliability and
stability of power network can be maintained by sensitive and selective relay sys-
tem. In optimum relay coordination problem, the total operating time of all relays
associated with the system is considered to minimize such that every relay can work

13
Optimal coordination of over current relay using opposition… 10727

with minimum time maintaining reliability of the system and a fast, stable and low
risk relay protection system can be achieved. Thus, the advantage of minimizing
total relay operating time is to maintain reliability, sensitivities and stability in
the power network. Commercially available relays like MiCOM P111, P115, P116
(overcurrent relay)are specified with operating time less than 10ms and P127, P521
(overcurrent relay) are specified with operating time less than 0.7ms [8, 20].

2.5 Opposition learning‑based gravitational search algorithm (OLGSA)

2.5.1 Basic gravitational search algorithm

Gravitational search algorithm (GSA) is a well known optimization algorithm based

on Newton’s law of gravitational force and interaction between objects of differ-
ent masses [17]. In this algorithm, the performance of the objects is measured by
masses of objects. The gravitational force creates an acceleration on all objects of
different masses globally, towards the heaviest mass of object. The object with high-
est mass indicated the global solution of the optimization problem. The position of
the objects along with object’s mass and gravitational coefficient is observed during
the updating part of objects performance in GSA. The position of an object indicates
the solution of the optimization problem. To determine the gravitational coefficient
and masses, a performance index is used. By proper updating the position, gravita-
tional mass and object masses, it leads the objects towards heaviest mass that will be
the optimum solution [17].

2.5.2 Steps to implement GSA

1. Initial populations are generated with random value for N number of objects. The
position of each object is described as follows:
( )
xj (t) = xj1 , xj2 , xj3 , … , xjd , xjN (7)

where x is potion, i is object number, d is dimension of N number of object.

2. The fitness of all objects in the population are calculated for an objective function.
The best and worst objects are selected as follows:
fiti (t) − worst(t)
mi (t) = , (8)
best(t) − worst(t)

m (t)
Mi (t) = ∑N i , (9)
j=1 j (t)
m

where fiti (t) is fitness value of ith object at time t, for minimization problem.
{ }
best(t) = min fitj (t) , (10)
j∈{1…N

13
10728 D. Acharya, D. K. Das

{
worst(t) = max fitj (t) (11)
j∈{1…N

3. Gravitational constant is updated as follows:

G = G0 exp (−𝛼t∕T). (12)
4. The force at time t due to object i on object j is given by
Mpi (t) × Maj (t) ( d )
Fijd (t) = G(t) xj (t) − xid (t) . (13)
Rij (t) + 𝜀

where Maj (t) is active mass of jth object, Mpi (t) is passive mass of ith object, G(t)
is gravitational coefficient at time t, Rij (t) is distance between ith and jth object.
5. Effective force is calculated as follows:
∑
Fid (t) = randj × Fijd (t)
j∈k
best (14)
j≠i

where kbest is the set of first kth objects.

6. According to the law of motion, the acceleration of object i due to force is
expressed as:
Fi (t)
ai (t) = (15)
Mii (t)

, where Mii (t) is inertial mass of ith object.

7. The velocity and position of object are updated as follows:
vi (t + 1) = randi × vi (t) + ai (t) (16)

Xi (t + 1) = Xi (t) + vi (t + 1). (17)

where rand is random number at ith iteration, v is velocity of object, a is the
acceleration of object.
8. Repeat step 2 to step 6 until the maximum iteration is completed.

2.5.3 Opposition learning concept

The concept of an opposition-based learning method was introduced in [26]. The

basic idea of this concept is to find optimum solution of a given object with the
help of randomly generated initial population and their opposite population in a
fixed search space. In this scheme, an objective function is run with an initially gen-
erated random population. If the fitness with random population is far from opti-
mum solution, the population from opposite direction can be considered to find the
best solution for the chosen objective function. In general, it is difficult to get best
solution with initial randomly generate population. There may be a chance that the

13
Optimal coordination of over current relay using opposition… 10729

population from opposite direction of initially generated population can provide best
solution. Therefore, the convergence may be accelerated by following iteratively
[26].

Definition 1 Consider a real number x in the range of [b1 , b2 ]. Then, the opposite
number xop is defined as follows:
xop = b1 + b2 − x. (18)
Let b1 = 0 and b2 = 1, then
xop = 1 − x. (19)
Definition 2 Consider h(x1 , x2 , x3 , … , xn ) be a point with x1 , x2 , x3 , … , xn as real
number in n dimensional space within the range of [bi , bi ] ∀i = 1, 2, … , n. Then, the
opposite number xopi is defined as follows:
xopi = b1i + b2i − xi ∀i[1,2,…,n] . (20)
Consider, h(x) is a function within the range of [b1 , b2 ] and h(xop ) is the opposite
learning function of h(x). If g(h(x)) be the solution of a chosen objective function for
h(x) and g(h(xop )) be the solution for h(xop ) then in each iteration, we proceed with
xop for g(h(xop )) < g(h(x)), otherwise proceed with x.

2.5.4 Opposition learning‑based gravitational search algorithm (OLGSA)

The concept of opposition learning scheme is merged with basic GSA to accelerate
the performance of convergence of the proposed optimization towards the optimum
solution. Generally, the population-based metaheuristic algorithms began with ran-
domly generated initial value in a specified range of objects or agents according to
the optimization problem. In GSA, the objects initial masses are also generated ran-
domly. Such initial solutions are used to find the global optimum. The initial objects
masses (randomly generated initial solution) in GSA are produced randomly using
expert knowledge like specified range of knowledge-based solution or different cri-
terion. As, the objects will work in search space there may be a chance where GSA
may not converge to global solution with such randomly generated initial solutions
[1]. Moreover, as the position updating in GSA of all objects is done by measur-
ing the distance between initial values and global values or solutions, the conver-
gence rate is also slow. To overcome such problem in GSA, the opposition learning
concept is applied. By incorporating the opposite direction populations of randomly
generated populations in the process of GSA, the solution can be made close to opti-
mum or global solution. Therefore, it leads the convergence rate of GSA.
Therefore, randomly generated initial population and their opposite population
based on the principle of opposition learning strategy are applied to the chosen
objective function. With the best solution among initial population and their oppo-
site direction population, the other steps of basic GSA are followed for next iteration

13
10730 D. Acharya, D. K. Das

Fig. 2  Opposition learning-

based gravitational search
algorithm (OLGSA)

13
Optimal coordination of over current relay using opposition… 10731

for further proceed. Thus, the global solution with fast convergence can be achieved
with proposed algorithm. The OLGSA algorithm is shown in Fig. 2.

2.5.5 OLGSA for relay coordination problem

In this section, the steps for implement OLGSA for relay coordination problem are
discussed.

Step 1 Initialize relay coordination problem in terms of population number within

relay operating time range, constants, maximum number of iteration, etc.
Step 2 Generate initial population for N number of population or relay with ran-
dom number in the required range and using (19) calculate the opposition popula-
tion.
Step 3 Evaluate objective function (4).
Step 4 Find the best and worst objects of initial population and its opposition
population.
Step 5 Compare the fitness of initial population and its opposition and with best
result follow step 6 and step 7.
Step 6 Update gravitational constant, mass (M), acceleration (a) for each relay
TDS.
Step 7 Update the velocity and position of objects.
Step 8 Do repetition of step 2 to step 6 till maximum iteration is reached.

3 Evaluation

3.1 Simulation and result analysis

In this section, the coordination among relay for single and multi mesh systems has
been presented. LabVIEW©2015 platform has been used to implement the OGSA
for relay coordination purpose. The main objective is to find optimal TDS value of
relay in a distribution system. The effectiveness of the OGSA for relay coordina-
tion problem has been verified with three types of power distribution network. The
details of the three network are taken from [14]. The performance of OGSA for such
coordination problem is compared with existing result for all cases.

3.1.1 Case study 1: relay coordination problem in single end multi parallel feeder
distribution system

In this section, the relay coordination in single end multi-parallel feeder distribution
system is considered with six OC relay as shown in Fig. 3. The line admittances
are assumed to be insignificant because of short distribution line. The line data for
single end multi parallel feeder distribution system is maintained in Table 2. The CT
ratios and plug settings for this case are indicated in Table 3. In Table 4, the details
of different fault locations, ai constant, and current sensing relay are presented. The
objective function can be written as follows:

13
10732 D. Acharya, D. K. Das

Fig. 3  Single end multi parallel feeder distribution system

Table 2  Line data for case study Line Impedance

1 [7]
1-2 0.08 + j1
1-3 0.16 + j2
2-3 0.08 + j1

Table 3  CT ratio and plug Relay CT ration Plug setting

settings of relays for case study
1 [7] R2 300:1 1
R4 600:1 1
R3 600:1 1
R6 600:1 1
R1 1000:1 1
R3 1000:1 1

Table 4  ai constant at different fault location for case study 1 for case study 1 [7]
Fault location Relay 1 Relay 2 Relay 3 Relay 4 Relay 5 Relay 6

K 75.91 – 75.91 – – 4.044

L 13.99 – 13.99 – 6.872 –
M 8.844 – 8.844 8.844 8.844 –
N 3.646 6.065 – 15.55 15.55 –

min z = 102.4x1 + 6.06x2 + 98.76x3 + 24.40x4 + 35.32x5 + 11.54x6 . (21)

Subject to the following operating time constant constraints:

– 3.646x1 ≥ 0.1,

13
Optimal coordination of over current relay using opposition… 10733

– 6.055x2 ≥ 0.1,
– 8.844x3 ≥ 0.1,
– 8.844x4 ≥ 0.1,
– 4.044x5 ≥ 0.1,
– 11.539x6 ≥ 0.1.

The range of TDS is 0.025 to 1.2. Therefore, the following correction has been con-
sidered as follows: x2 , x3, x4 and x5 are greater than or equal to 0.025. Considering
CTI criterion, the following relationships are considered as follows:

– 15.55x4 − 6.065x2 ≥ 0.3,

– 8.844x1 − 8.844x3 ≥ 0.3,
– 4.844x5 − 8.844x4 ≥ 0.3,
– 75.91x3 − 11.53x6 ≥ 0.3,
– 13.998x1 − 13.998x3 ≥ 0.3.

3.1.2 Simulation analysis for case study 1

In this case study, a single end loop system consisting of six OC relay has been
simulated in LabVIEW©2015 platform. For this purpose, the range of TDS is con-
sidered as 0.025 to 1.2 sec. OLGSA is implemented in LabVIEW©2015 platform.
The OLGSA has been run for 1000 initial population and 1500 iterations. The value
of optimum TDS for all relays and operating time are presented in Table 5.

3.1.3 Discussion for case study 1

The relay coordination of single end loop distribution system has been performed
with OLGSA in LabVIEW©2015 platform. The results have been presented in
Table 5. By analysing Table 5, it is found that the minimum operating time for the
system is improved by 7.29 s, 6.62 s, 5.43 s, 2.97 s and 2.9 s in comparison with

Table 5  Simulation result for case study 1: TDS and total operating time
Approaches Firefly Continuous Chaotic firefly Root tree Water cycle Proposed
algorithm genetic algo- algorithm [7] algorithm algorithm [14] OLGSA
[6] rithm [15] [27]

TDS 1 0.027 0.0765 30.027 0.059 0.0589 0.0265

TDS 2 0.13 0.034 0.221 0.025 0.025 0.0250
TDS 3 0.025 0.0339 0.025 0.025 0.025 0.0256
TDS 4 0.025 0.036 0.025 0.029 0.029 0.0250
TDS 5 0.025 0.0711 0.029 0.065 0.0626 0.0554
TDS 6 0.489 0.0294 0.363 0.025 0.025 0.0857
Total operat- 16.25 15.58 14.39 11.93 11.86 8.96
ing time in
second

13
10734 D. Acharya, D. K. Das

Firefly algorithm [6], Continuous genetic algorithm [15], Chaotic firefly algorithm
[7], Root tree algorithm [27] and Water cycle algorithm [14], respectively.

3.1.4 Case Study 2: Relay Coordination Problem in Multi Loop Single end Fed
Distribution System

In this section, the relay coordination in multi-loop single end fed Distribution Sys-
tem is considered. It consists of eight over current relay as shown in Fig. 4. The
line admittances are assumed to be insignificant because of short distribution line.
The data for CT ratios and plug settings are maintained in Table 6. The relation
between primary and backup relay under fault condition at various points is given
in Table 7. In Table 8, the details of different fault locations, ai constant, and current
sensing relay are presented. In this system, eight constraints are considered for eight
OC relay with minimum relay operating time and five CTI constraints are chosen
as coordination criteria. The minimum operating time for all relay is considered as

Fig. 4  Multi loop single end fed distribution system

Table 6  CT ratio and plug Relay CT ration Plug setting

settings of relays [14]
R1 1000 : 1 0.8
R2 1000 : 1 0.8
R3 500 : 1 0.8
R4 500 : 1 0.8
R5 1000 : 1 0.8
R6 1000 : 1 0.8
R7 500 : 1 0.5

Table 7  Fault location and Fault location Primary relay Backup relay Fault current(A)
respective primary, backup
relay and fault current for case K 5, 6 -, 3 3289, 1096.5
study-2 [14]
L 7 3,5 1315.8
M 3, 4 1, 5 2193, 1315.5
N 1, 2 -, 4 6759, 939

13
Optimal coordination of over current relay using opposition… 10735

Table 8  ai constant at different fault location for case study-2 [14]
Fault location Relay 1 Relay 2 Relay 3 Relay 4 Relay 5 Relay 6 Relay 7

K 10.035 10.035 2.971 4.9804 – – 4.9804

L 5.584 5.584 4.227 3.551 – – –
M 4.281 4.281 4.9804 2.971 – – –
N 2.267 3.837 10.035 – 3.297 – –
O 3.837 2.267 10.035 – – 3.297 –
P 2.971 2.971 5.749 – – – –

0.1s and CTI is considered as 0.6s. The TDS of all relays are denoted as xi where
i = 1, 2, … , 7.
The objective function can be written as follows:
min z =10.24x1 + 43.77x2 + 16.725x3 + 13.97x4 + 24.63x5 + 22.16x6
(22)
+ 4.145x7

Subject to the following operating time constant constraints:

– 2.97x1 ≥ 0.1,
– 2.97x2 ≥ 0.1,
– 5.584x3 ≥ 0.1,
– 4.980x4 ≥ 0.1,
– 3.297x5 ≥ 0.1,
– 3.297x6 ≥ 0.1,
– 4.980x7 ≥ 0.1,
– 10.035x8 ≥ 0.1.

within the range of TDS, the following criteria have been considered:

– x3 ≥ 0.025,
– x4 ≥ 0.025,
– x7 ≥ 0.025,
– x8 ≥ 0.025.

Considering CTI criteria, the following relationships are derived as follows:

– 5.749x3 − 5.749x8 ≥ 0.6,

– 5.584x1 − 3.551x4 ≥ 0.6,
– 5.584x2 − 3.551x4 ≥ 0.6,
– 4.980x4 − 4.980x7 ≥ 0.6,
– 4.281x1 − 2.971x4 ≥ 0.6,
– 4.980x3 − 4.980x8 ≥ 0.6,
– 10.035x8 − 3.297x5 ≥ 0.6,
– 10.035x8 − 3.297x6 ≥ 0.6.

13
10736 D. Acharya, D. K. Das

3.1.5 Simulation analysis for case study 2

In this case study, a multi loop single fed distribution with eight relay has been simu-
lated in LabVIEW©2015 platform. For this purpose, the range of TDS is considered
as 0.025 to 1.2. For this system, OLGSA is implemented in LabVIEW©2015 plat-
form. The OLGSA has been run for 1000 initial population and 1500 iterations. The
values of optimum TDS for all relays and operating time are presented in Table 9.

3.1.6 Discussion for case study 2

The relay coordination optimization for multi loop single fed distribution system has
been performed with OLGSA in LabVIEW©2015. The results have been presented
in Table 9. By analysing Table 9, it is found that the minimum operating time for the
system is improved by 11.50 sec and 0.97 sec in comparison with the genetic algo-
rithm [4] and root tree algorithm [27], respectively.

3.1.7 Case study 3: relay coordination problem in single loop distribution system

In this section, the relay coordination in single loop distribution system is consid-
ered as shown in Fig. 5. It is fed from one end and it consists of seven over current
relay. The line admittances are assumed to be insignificant because of short distribu-
tion line. All relays are directional relay except relay 1 and relay 5 (non-directional).
The data for CT ratios and plug settings are presented in Table 10. The relation
between primary and back up relay under fault condition at various points is given in
Table 11. In Table 12, the details of different fault locations, ai constant, and current
sensing relay are presented.
In this system, seven constraints are considered for seven relay with minimum
relay operating time and five constraints are chosen as coordination criteria. The

Table 9  Simulation result for Time dial setting Genetic Root tree Proposed OLGSA
case study 2: TDS and total algorithm algorithm
operating time [4] [27]

TDS 1 0.297 0.252 0.084

TDS 2 0.297 0.252 0.026
TDS 3 0.227 0.20 0.183
TDS 4 0.173 0.151 0.0264
TDS 5 0.06 0.03 0.0251
TDS 6 0.06 0.03 0.025
TDS 7 0.04 0.025 0.028
TDS 8 0.113 0.08 0.025
Total operating 15.70 5.17 4.20
time in second

13
Optimal coordination of over current relay using opposition… 10737

Fig. 5  Single Loop Distribution System

Table 10  CT ratio and plug Relay CT ration Plug setting

settings of relays for case study
3 [14] R1 1000:1 0.8
R2 1000:1 0.8
R3 500:1 0.8
R4 500:1 0.8
R5 1000:1 0.8
R6 1000:1 0.8
R7 500:1 0.5

Table 11  Fault location and Fault location Primary relay Backup relay Fault current(A)
respective primary, backup relay
and fault current for case study K 5, 6 –, 3 3289, 1096.5
3 [14]
L 7 3,5 1315.8
M 3, 4 1, 5 2193, 1315.5
N 1, 2 –, 4 6759, 939

Table 12  ai constant at different fault location for case study 3 [14]
Relay K (Fault location) L (Fault location) M (Fault location) N (Fault location)

R1 – – 3.252 6.872
R2 – – 43.776 –
R3 6.872 5.809 – 4.044
R4 – – 8.159 5.811
R5 4.881 –5.809 – 14.01
R6 22.165 – – –
R7 – 4.145 – -

13
10738 D. Acharya, D. K. Das

minimum operating time for all relays is considered as 0.1s and CTI is considered as
0.2s. The relay 1 will work for fault at location M and N as shown in Fig. 5. Math-
ematically, it can be written as follows:
a1 = 3.252 + 6.872 = 10.124, (23)
Relay 2 work for fault at point M. Mathematically, it is described as follows:
a2 = 43.776, (24)
Relay 3 can detect fault at fault locations K, L and N and can be written as follows:
a3 = 6.872 + 5.809 + 4.044 = 16.725, (25)
Relay 4 can detect fault at fault locations M, N and can be written as follows:
a4 = 8.159 + 5.811 = 13.97, (26)
Relay 5 can detect fault at fault locations K, L and N and can be written as follows:
a5 = 4.811 + 5.809 + 14.1 = 24.63, (27)
Relay 6 can detect fault at fault locations K and can be written as follows:
a6 = 22.165, (28)
Relay 7 can detect fault at fault locations L and can be written as follows:
a7 = 4.145, (29)
Therefore, the objective function cab be written as follows:
min z =10.24x1 + 43.77x2 + 16.725x3 + 13.97x4 + 24.63x5 + 22.16x6
(30)
+ 4.145x7 ,

where xi is TDSi of Relayi and i = 1, 2 … , 7.

Therefore, the constrains are as follows:

– 3.255x1 ≥ 0.1,
– 43.77x2 ≥ 0.1,
– 4.044x3 ≥ 0.1,
– 5.811x4 ≥ 0.1,
– 4.881x5 ≥ 0.1,
– 22.165x6 ≥ 0.1,
– 4.14x7 ≥ 0.1.

In terms of minimum value of TDS, the constraints are chosen as follows: xi ≥ 0.025
i = 1, 2, … , 7.
Considering CTI criteria, the following relationships are considered as follows:

– 43.77x2 − 8.159x4 ≥ 0.2,

13
Optimal coordination of over current relay using opposition… 10739

– 6.872x1 − 4.044x3 ≥ 0.2,

– 14.01x5 − 5.811x4 ≥ 0.2,
– 6.872x3 − 22.165x6 ≥ 0.2,
– 5.809x3 − 13.998x7 ≥ 0.2.

3.1.8 Simulation analysis for case study 3

The relay coordination optimization is simulated in LabVIEW©2015 platform. For

the simulation purpose, the range of TDS is considered as 0.025 to 1.2. OLGSA
is implemented in LabVIEW©2015 platform. The proposed algorithm is run with
1500 initial population for 20000 iterations. The values of optimum TDS for all
relays and operating time are presented in Table 13.

3.1.9 Discussion for case study 3

The relay coordination optimization for single loop distribution system has been
performed with OLGSA in LabVIEW©2015. The results have been presented in
Table 13. By analysing Table 13, the minimum operating time is achieved 6.05 sec.
Therefore, the performance of OLGSA for relay coordination problem of single loop
distribution system is better than Simplex method by 9.65 sec [3].

4 Conclusion

Opposition learning-based gravitational search algorithm is proposed for optimiza-

tion of over current relay coordination in distribution system. The total operating
time of relay system is minimized by the proposed method. The optimum value of
time dial setting (TDS) is found for a fixed plug setting (PS) under the constraints of
coordination time interval (CTI). The proposed approach has been verified for relay
coordination problem with three different distribution structures (single end multi-
parallel feeder distribution system, multi loop single end fed distribution system, sin-
gle loop distribution system). The simulation results for all three systems are com-
pared with firefly algorithm, continuous genetic algorithm, chaotic firefly algorithm,

Table 13  Simulation result for Time dial setting Simplex method [3] Proposed OLGSA
case study 3: TDS and total
operating time TDS 1 0.238 0.026
TDS 2 0.12 0.025
TDS 3 0.36 0.055
TDS 4 0.031 0.087
TDS 5 0.027 0.153
TDS 6 0.025 0.025
TDS 7 0.08 0.036
Total operating time in 15.70 6.05
second

13
10740 D. Acharya, D. K. Das

Root tree algorithm and Water cycle algorithm. The analysis of simulation results
shows that the minimum operating time, optimum TDS under constraints of CTI is
achieved with applying OLGSA for all three distribution structures. The minimized
operating time under all constraints are suggested to avoid the damage due to fault
for improper operation of circuit breaker. Therefore, the reliability of the power sys-
tem is improved with optimum coordination of relay with proposed approach.

Declarations

Conflict of interest The authors declare that they have no conflict of interest.

References
1. Abd Elaziz M, Oliva D, Xiong S (2017) An improved opposition-based sine cosine algorithm for
global optimization. Exp Syst Appl 90:484–500
2. Alkaran DS, Vatani M, Sanjari M, Gharehpetian GB, Naderi MS (2016) Optimal overcurrent relay
coordination in interconnected networks by using fuzzy-based ga method. IEEE Trans Smart Grid
9(4):3091–3101
3. Bedekar PP, Bhide SR (2010) Optimum coordination of directional overcurrent relays using the
hybrid ga-nlp approach. IEEE Trans Power Deliv 26(1):109–119
4. Bedekar PP, Bhide SR (2011) Optimum coordination of overcurrent relay timing using continuous
genetic algorithm. Exp Syst Appl 38(9):11286–11292
5. Elsadd MA, Kawady TA, Taalab AMI, Elkalashy NI (2021) Adaptive optimum coordination of
overcurrent relays for deregulated distribution system considering parallel feeders. Electrical Engi-
neering, 1–19
6. Gokhale S, Kale V (2014) Application of the firefly algorithm to optimal over-current relay coordi-
nation. In: 2014 International Conference on Optimization of Electrical and Electronic Equipment
(OPTIM). IEEE, pp 150–154
7. Gokhale S, Kale V (2016) An application of a tent map initiated chaotic firefly algorithm for opti-
mal overcurrent relay coordination. Int J Electr Power Energy Syst 78:336–342
8. Hemmati R, Mehrjerdi H (2019) Non-standard characteristic of overcurrent relay for minimum
operating time and maximum protection level. Simul Modelling Pract Theory 97(101):953
9. Kalage AA, Ghawghawe ND (2016) Optimum coordination of directional overcurrent relays using
modified adaptive teaching learning based optimization algorithm. Intell Ind Syst 2(1):55–71
10. Khurshaid T, Wadood A, Farkoush SG, Kim CH, Cho N, Rhee SB (2019) Modified particle swarm
optimizer as optimization of time dial settings for coordination of directional overcurrent relay. J
Electr Eng Technol 14(1):55–68
11. Khurshaid T, Wadood A, Farkoush SG, Kim CH, Yu J, Rhee SB (2019) Improved firefly algorithm
for the optimal coordination of directional overcurrent relays. IEEE Access 7:78503–78514
12. Korashy A, Kamel S, Jurado F, Youssef AR (2019) Hybrid whale optimization algorithm and grey
wolf optimizer algorithm for optimal coordination of direction overcurrent relays. Electric Power
Compon Syst 47(6–7):644–658
13. Korashy A, Kamel S, Nasrat L, Jurado F (2020) Developed multi-objective grey wolf optimizer with
fuzzy logic decision-making tool for direction overcurrent relays coordination. Soft Comput 1–13
14. Kudkelwar S, Sarkar D (2019) Online implementation of time augmentation of over current relay
coordination using water cycle algorithm. SN Appl Sci 1(12):1628
15. Moirangthem J, Krishnanand K, Dash SS, Ramaswami R (2013) Adaptive differential evolution
algorithm for solving non-linear coordination problem of directional overcurrent relays. IET Gener
Transm Distrib 7(4):329–336
16. Rajput VN, Pandya KS (2017) Coordination of directional overcurrent relays in the interconnected
power systems using effective tuning of harmony search algorithm. Sust Comput Info Syst 15:1–15

13
Optimal coordination of over current relay using opposition… 10741

17. Rashedi E, Nezamabadi-Pour H, Saryazdi S (2009) Gsa: a gravitational search algorithm. Inf Sci
179(13):2232–2248
18. Saha D, Datta A, Das P (2016) Optimal coordination of directional overcurrent relays in power
systems using symbiotic organism search optimisation technique. IET Gener Transm Distrib
10(11):2681–2688
19. Sarwagya K, Nayak PK, Ranjan S (2020) Optimal coordination of directional overcurrent relays in
complex distribution networks using sine cosine algorithm. Electric Power Syst Res 187(106):435
20. Schneider E (2015) Micom overcurrent relay-technical guide. https://​www.​se.​com/​in/​en/​produ​ct-​
range-​downl​oad/​60758-​micom-​p12x/#/​docum​ents-​tab
21. Shahzad F, Baig AR, Masood S, Kamran M, Naveed N (2009) Opposition-based particle swarm
optimization with velocity clamping (ovcpso). In: Advanced Computational Intelligence. Springer,
pp 339–348
22. Shih MY, Conde A, Ángeles-Camacho C (2019) Enhanced self-adaptive differential evolution
multi-objective algorithm for coordination of directional overcurrent relays contemplating maxi-
mum and minimum fault points. IET Gener Transm Distrib 13(21):4842–4852
23. Shih MY, Salazar CAC, Enríquez AC (2015) Adaptive directional overcurrent relay coordination
using ant colony optimisation. IET Gener Transm Distrib 9(14):2040–2049
24. Srinivas S, Swarup KS (2019) A new mixed integer linear programming formulation for protection
relay coordination using disjunctive inequalities. IEEE Power Energy Technol Syst J 6(2):104–112
25. Swathika OG, Hemamalini S (2020) Graph theory and optimization algorithms aided adaptive pro-
tection in reconfigurable microgrid. J Electr Eng Technol 15(1):421–431
26. Tizhoosh HR (2005) Opposition-based learning: a new scheme for machine intelligence. In: Interna-
tional Conference on Computational Intelligence for Modelling, Control and Automation and Inter-
national Conference on Intelligent Agents, Web Technologies and Internet Commerce (CIMCA-
IAWTIC’06), vol 1. IEEE, pp 695–701
27. Wadood A, Gholami Farkoush S, Khurshaid T, Kim CH, Yu J, Geem ZW, Rhee SB (2018) An
optimized protection coordination scheme for the optimal coordination of overcurrent relays using a
nature-inspired root tree algorithm. Appl Sci 8(9):1664

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published
maps and institutional affiliations.

13