Comparative analysis of directional overcurrent

relay coordination using linear and nonlinear

protection schemes
2023 International Conference on Computer, Electronics & Electrical Engineering & their Applications (IC2E3) | 979-8-3503-3800-3/23/$31.00 ©2023 IEEE | DOI: 10.1109/IC2E357697.2023.10262675

Raghvendra Tiwari Ravindra Kumar Singh Niraj Kumar Choudhary

Electrical Engineering Department Electrical Engineering Department Electrical Engineering Department
MNNIT Allahabad MNNIT Allahabad MNNIT Allahabad
Prayagraj, INDIA Prayagraj, INDIA Prayagraj, INDIA
[email protected] [email protected] [email protected]

Abstract— In this paper, the relay coordination problem is determines the final TMS and PS [3]. A graph theory
formulated as a linear and non-linear optimization problem approach is used for the selection of breakpoint relays [4]. For
with a number of operational constraints. The primary objective the relay coordination problem, traditional optimization
of DOCR coordination is to identify the optimal relay settings to techniques were employed in the late 1980s. An overcurrent
achieve the least overall operation time of all relays without any relay's operating time is determined by a few relay
relay pair miscoordination. Also, a comparative analysis of relay
parameters, including time multiplier settings (TMS), plug
coordination has been performed using standard and user-
defined relay characteristics. In this study, the application of setting (PS), and relay characteristics coefficients (α, β) [5].
grey wolf optimization (GWO) for the optimal coordination of As the operating time of an overcurrent relay is a linear
DOCRs in a mesh power system is discussed. Additionally, in function of the time dial setting, the linear programming
terms of optimization algorithm, the obtained results are also approach is only useful for improving the time dial settings
compared with a genetic algorithm (GA). The performance of [6]. For OCR coordination, the researchers provide a variety
the proposed algorithm is tested on the distribution part of a of LPP approaches [7-8].In contrast, non-linear optimization
modified IEEE-14 bus test system. was used to find the optimal relay settings for both TMS and
PS. However, NLP techniques take a lot of time and are
Keywords—overcurrent relay coordination, coordination time complicated. The DOCR coordination problem is sometimes
margin, GWO, GA treated as a linear programming problem (LPP) in order to
avoid the complexity of the NLP approaches [9]. In NLP
I. INTRODUCTION approach, the optimal relay settings are obtained, by
The protection system is playing a crucial role in ensuring considering both TMS and PS as a continuous decision
that customers will always have access to electricity as a variables. In literature, the optimal overcurrent relay settings
result of the expansion of power system networks, which are obtained via NLP approach using chaotic cuckoo search
have mostly become interconnected and have complex (CCS) [10], water cycle algorithm (WCA) [11], teaching
structures. One of the protective relays that is widely used in learning based optimization (TLBO) [12], human behavior
the protection of distribution systems is the directional based optimization (HBBO) [13] and many more. In this
overcurrent relay (DOCR) [1]. They are also regarded as an paper, the GWO algorithm is utilized to optimize the values
effective and affordable technique for protecting the sub- of TMS, PS α, and β. The magnitude of relay characteristics
transmission and mesh distribution electrical systems. The coefficient for normal inverse (NI), very inverse (VI) and
protective relays in the power system networks must be extremely inverse (EI) are shown in Table I [14]. The relay
coordinated with one another in order to effectively protect coordination problem is formulated via two different
the power system and maintain service across the healthy approaches namely linear, and non-linear. A comprehensive
part. The relay coordination process is defined as determining discussion of comparative analysis between linear and non-
the optimal relay settings within the constraints specified. linear programming approaches for relay coordination
When the main relays efficiently clear the fault in the primary problem in this paper can be considered as novelty of the
protected zone, it is possible to achieve the optimal level of proposed work. Also the optimal relay settings are obtained
coordination for the DOCRs. However, if the main relay fails using grey wolf optimization (GWO) and genetic algorithm
to clear the fault, the backup relay must operate to clear the (GA). All considered approaches are tested on the
same fault after a predetermined time interval known as the distribution part of the modified IEEE 14 bus test system.
coordination time interval (CTI) [2].
Numerous optimization strategies have recently come into TABLE I. RELAY CHARACTERISTICS COEFFICIENT AS PER IEC-60255
STANDARD
existence for the best coordination of DOCRs. These are
broadly classified into two categories namely traditional
approaches and meta-heuristic approaches. In the traditional Relay characteristics α β
method, fault analysis is first carried out, and then the meshed NI 0.14 0.02
VI 13.5 1.0
network is converted into radial segments. The relay at the far
EI 80 2.0
end is then set first, and a backup relay is set immediately
after. This process is repeated until all relays have been
The remaining sections of this research report are
considered. This procedure involves a series of iterations. The
organised as follows: The mathematical formulation of
selection of initial relays, also known as break point relays,
DOCR coordination problems is discussed in Section II, and

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

Authorized licensed use limited to: Motilal Nehru National Institute of Technology. Downloaded on September 07,2025 at 05:30:20 UTC from IEEE Xplore. Restrictions apply.
 the test system simulation and descriptions are explained in R14 R16 11.04 3.0125
Section III. Section IV presents the results and discussions of R15 R3 17.728 2.031
L8 R15 R13 17.728 7.78
the modified IEEE 14-bus test system. The major R16 R5 9.734 5.71
contribution of this paper is concluded in Section V.
III. PROBLEM FORMULATION
II. TEST CASE DESCRIPTION
In this paper, the relay coordination problem is formulated
In this paper, the proposed overcurrent protection using linear and nonlinear programming approaches. The
technique is investigated on the distribution part of a generalized mathematical expression of an objective function
modified IEEE-14 bus test system, as shown in figure 1. The
test system consists of eight lines, seven buses, and three is shown in (1). In (1), woci is a non-negative weight
DGs. To protect the complete test system, two DOCRs are associated with primary relay operating time. The value of
placed at the end of each line. Therefore, a total of 16 DOCRs woci is fixed for all DOCRs and taken as unity. The operating
are deployed to protect the complete test system. In the time of ith primary DOCR is expressed as toci shown in (2). In
context of DG integration, two IBDGs of 20 MVA capacity
each are connected at bus B2, B7 and one SBDG of 50 MVA (2), PSi , TMSi , αi and βi are plug setting, time multiplier
capacity is connected at bus B1. The utility grid is connected settings, and relay characteristic coefficients of the ith DOCR.
16
through buses B3 and B6. The CT ratios of all considered OF = min  woci × toci (1)
DOCRs are given in Table II. The data for the three i =1
symmetrical mid-point fault currents are taken from [14]. All Where,
other test system information can be taken from [15]. α i × TMS i
toci = βi
(2)
 If 
IBDG SG IBDG  i  −1
 PS × CTR 
B7 B1 B2

R16 R6 L3 R5 R3 R1 L1 R2 R7
CTR = Current transformer ratio
If = Fault current flowing through relay
A. Constraint set to maintain CTI between primary and
L2 corresponding backup relay pairs
L8 L4
To avoid the mal-operation between relay pairs, it is
recommended to maintain a predefined time gap between the
L7 L6 L5
primary and backup relay operating time. If toci and tocj are the
R15 R4 R14 R13 R12 R11 R10 R9 R8 operating time of the primary (Ri) and corresponding backup
relay (Rj) for a fault location. The mathematical expression
B6 B5 B4 B3
for relay coordination constraint to maintain CTI is shown in
(3).

tocj − toci ≥ CTI (3)

Fig 1 Single line diagram of distribution part of IEEE-14 bus test system B. Constraint set to select optimal values of decision
variables between minimum and maximum values
TABLE II. CURRENT FLOWS THROUGH RELAYS WITH The optimal values of decision variables in relay
PRIMARY/BACKUP RELAY PAIRS AT VARIOUS FAULT POINTS
coordination problem must lies between the minimum and
Faulty Primary Backup Fault Current Through maximum values as designed by the relay manufacturer. In
Line Relay Relay Relay Coil (A)
this study the desired range of TMS and PS are considered as
If,Pri If,Backup
(0.1-1.1) and (0.5-2.0) [14]. Similarly, the operating range of
R1 R4 12.075 4.785
α and β are taken as (0.14-80) and (0.02-2.00) respectively
L1 R1 R6 12.075 4.06
R2 R8 17.175 6.465 [5].The following constraints from (4) to (7) can be
R3 R2 9.561 8.205 considered to limit the optimal values between lower and
L2 R3 R6 9.561 3.47 upper bounds:
R4 R13 16.075 9.918
R4 R16 16.075 1.696 i
R5 R2 17.196 6.54
TMSmin ≤ TMS i ≤ TMSmax
i
(4)
L3 R5 R4 17.196 4.00 i i i
PS min ≤ PS ≤ PS max (5)
R6 R15 16.785 6.187 i i i
L4 R7 R1 7.038 4.97 α min ≤α ≤α max (6)
R8 R10 16.793 4.875 i i i
L5 R9 R7 16.038 3.896
β min ≤β ≤β max (7)
R10 R12 11.634 5.214
L6 R11 R9 19.90 9.296 Where,
R12 R14 7.18 5.906 i
R13 R11 18.28 11.99
TMS min Minimum value of TMS of relay Ri
i
L7 R14 R3 11.04 3.676 TMS max Maximum value of TMS of relay Ri

Authorized licensed use limited to: Motilal Nehru National Institute of Technology. Downloaded on September 07,2025 at 05:30:20 UTC from IEEE Xplore. Restrictions apply.
 i
PSmin Minimum value of PS of relay Ri α*x(13)*(1/((9.919)^β-1))-α*x(4)*(1/((16.075)^β-1)) ≥ 0.2
i α*x(16)*(1/((1.697)^β-1))-α*x(4)*(1/((16.075)^β-1)) ≥ 0.2
PSmax Maximum value of PS of relay Ri
α*x(2)*(1/((6.540)^β-1))-α*x(5)*(1/((17.197)^β-1)) ≥ 0.2
i
α min Minimum value of α of relay Ri α*x(4)*(1/((4.00)^β-1))-α*x(5)*(1/((17.197)^β-1)) ≥ 0.2
i
α max Maximum value of α of relay Ri α*x(15)*(1/((6.188)^β-1))-α*x(6)*(1/((16.785)^β-1)) ≥ 0.2
i α*x(1)*(1/((4.970)^β-1))-α*x(7)*(1/((7.038)^β-1)) ≥ 0.2
β min Minimum value of β of relay Ri
α*x(10)*(1/((4.875)^β-1))-α*x(8)*(1/((16.794)^β-1)) ≥ 0.2
i
β max Maximum value of β of relay Ri α*x(7)*(1/((3.896)^β-1))-α*x(9)*(1/((16.038)^β-1)) ≥ 0.2
α*x(12)*(1/((5.214)^β-1))-α*x(10)*(1/((11.633)^β-1)) ≥ 0.2 (10)
C. Constraint set related to primary relay operating time
α*x(9)*(1/((9.296)^β-1))-α*x(11)*(1/((19.904)^β-1)) ≥ 0.2
In order to protect the relay coordination from the α*x(14)*(1/((5.907)^β-1))-α*x(12)*(1/((7.180)^β-1)) ≥ 0.2
temporary fault, it is recommended that the primary relay α*x(11)*(1/((11.996)^β-1))-α*x(13)*(1/((18.288)^β-1) ≥ 0.2
operate with some minimal delay. In this context, the primary
α*x(3)*(1/((3.677)^β-1))-α*x(14)*(1/((11.043)^β-1)) ≥ 0.2
relay must operate between the minimum and maximum
α*x(16)*(1/((3.013)^β-1))-α*x(14)*(1/((11.043)^β-1) ≥ 0.2
values. If toci ,min and toci ,max are the minimum and maximum α*x(3)*(1/((2.032)^β-1))-α*x(15)*(1/((17.728)^β-1)) ≥ 0.2
relay operating time, the mathematical expression to limit the α*x(13)*(1/((7.781)^β-1))-α*x(15)*(1/((17.728)^β-1) ≥ 0.2
relay operating time is given by (8). The minimum and α*x(5)*(1/((5.716)^β-1))-α*x(16)*(1/((9.734)^β-1)) ≥ 0.2
maximum value of relay operating time is considered as 0.1
and 4.0s [16]. The relay constraints associated with minimum primary relay
operating time are as follows:
toci ,max ≥ toci ≥ toci ,min (8)
α*x(1)*(1/((12.075)^β-1)) ≥ 0.1
IV. IMPLEMENTATION OF PROPOSED PROTECTION SCHEME α*x(2)*(1/((17.175)^β-1)) ≥ 0.1
α*x(3)*(1/((9.562)^β-1)) ≥ 0.1
In this section the mathematical formulation of objective α*x(4)*(1/((16.075)^β-1)) ≥ 0.1
function and the relay constraints related to the relay α*x(5)*(1/((17.197)^β-1)) ≥ 0.1
coordination problem is discussed. α*x(6)*(1/((16.785)^β-1)) ≥ 0.1
A. Linear programming approach α*x(7)*(1/((7.038)^β-1)) ≥ 0.1
α*x(8)*(1/((16.794)^β-1)) ≥ 0.1
In the linear programming approach, the PS value is kept (11)
α*x(9)*(1/((16.038)^β-1)) ≥ 0.1
fixed, and the operating time of DOCR is directly
α*x(10)*(1/((11.633)^β-1)) ≥ 0.1
proportional to TMS only. The value of relay characteristics
α*x(11)*(1/((19.904)^β-1)) ≥ 0.1
coefficients are fixed according to NI, VI and EI. In this
approach, the decision variables x(1)-x(16) are associated α*x(12)*(1/((7.180)^β-1)) ≥ 0.1
with TMS1-TMS16 of each DOCR. The objective function of α*x(13)*(1/((18.288)^β-1)) ≥ 0.1
the relay coordination problem using a linear programming α*x(14)*(1/((11.043)^β-1)) ≥ 0.1
approach for the distribution part of the IEEE-14 bus system α*x(15)*(1/((17.728)^β-1)) ≥ 0.1
is demonstrated as follows: α*x(16)*(1/((9.734)^β-1)) ≥ 0.1

Min f= B. Non-Linear programming approach

α*x(1)*(1/((12.075)^β-1))+α*x(2)*(1/((17.175)^β-1)) In this approach, the PS is also considered a continuous
+α*x(3)*(1/((9.562)^β-1))+α*x(4)*(1/((16.075)^β-1)) decision variable along with the TMS. In this approach, the
+α*x(5)*(1/((17.197)^β-1))+α*x(6)*(1/((16.785)^β-1))
decision variables x(1)-x(16) are associated with TMS1-
+α*x(7)*(1/((7.038)^β-1))+α*x(8)*(1/((16.794)^β-1))
+α*x(9)*(1/((16.038)^β-1))+α*x(10)*(1/((11.633)^β-1)) (9) TMS16, and x(17)-x(32) are associated with PS1-PS16. The
+α*x(11)*(1/((19.904)^β-1))+α*x(12)*(1/((7.180)^β-1)) value of relay characteristics coefficients are fixed according
+α*x(13)*(1/((18.288)^β-1))+α*x(14)*(1/((11.043)^β-1)) to NI, VI and EI. The objective function of the relay
+α*x(15)*(1/((17.728)^β-1))+α*x(16)*(1/((9.734)^β-1)) coordination problem using a non-linear programming
approach for the distribution part of the IEEE-14 bus system
Relay constraints are classified into two types. The first is demonstrated as follows:
kind of relay constraints are associated with primary and
corresponding backup relay pairs, and another kind of relay Min f=α*x(1)*(1/((12.075/x(17))^β-1))
constraint is associated with the operating time of primary +α*x(2)*(1/((17.175/x(18))^β-1))
DOCRs. The constraints associated with the primary-backup +α*x(3)*(1/((9.562/x(19))^β-1))
+α*x(4)*(1/((16.075/x(20))^β-1))
relay coordination with CTI taken as 0.2 are as follows:
+α*x(5)*(1/((17.197/x(21))^β-1))
+α*x(6)*(1/((16.785/x(22))^β-1))
α*x(4)*(1/((4.785)^β-1))-α*x(1)*(1/((12.075)^β-1)) ≥ 0.2 (12)
+α*x(7)*(1/((7.038/x(23))^β-1))
α*x(6)*(1/((4.060)^β-1))-α*x(1)*(1/((12.075)^β-1)) ≥ 0.2 +α*x(8)*(1/((16.794/x(24))^β-1))
α*x(8)*(1/((6.466)^β-1))-α*x(2)*(1/((17.175)^β-1)) ≥ 0.2 (10) +α*x(9)*(1/((16.038/x(25))^β-1))
α*x(2)*(1/((8.205)^β-1))-α*x(3)*(1/((9.562)^β-1)) ≥ 0.2 +α*x(10)*(1/((11.633/x(26))^β-1))
+α*x(11)*(1/((19.904/x(27))^β-1))
α*x(6)*(1/((3.470)^β-1))-α*x(3)*(1/((9.562)^β-1)) ≥ 0.2
+α*x(12)*(1/((7.180/x(28))^β-1))

Authorized licensed use limited to: Motilal Nehru National Institute of Technology. Downloaded on September 07,2025 at 05:30:20 UTC from IEEE Xplore. Restrictions apply.
 +α*x(13)*(1/((18.288/x(29))^β-1)) α*x(3)*(1/((2.032/x(19))^β-1))
+α*x(14)*(1/((11.043/x(30))^β-1)) -α*x(15)*(1/((17.728/x(31))^β-1)) ≥ 0.2
+α*x(15)*(1/((17.728/x(31))^β-1))
+α*x(16)*(1/((9.734/x(32))^β-1)) α*x(13)*(1/((7.781/x(29))^β-1)) (13)
-α*x(15)*(1/((17.728/x(31))^β-1)) ≥ 0.2
The constraints associated to the primary-backup relay
coordination with CTI taken as 0.2 are α*x(5)*(1/((5.716/x(21))^β-1))
-α*x(16)*(1/((9.734/x(32))^β-1)) ≥ 0.2
α*x(4)*(1/((4.785/x(20))^β-1))
-α*x(1)*(1/((12.075/x(17))^β-1)) ≥ 0.2 The relay constraints associated with minimum primary
relay operating time are as follows:
α*x(6)*(1/((4.060/x(22))^β-1))
-α*x(1)*(1/((12.075/x(17))^β-1)) ≥ 0.2 α*x(1)*(1/((12.075/x(17))^β-1)) ≥ 0.1
α*x(2)*(1/((17.175/x(18))^β-1)) ≥ 0.1
α*x(8)*(1/((6.466/x(24))^β-1))
α*x(3)*(1/((9.562/x(19))^β-1)) ≥ 0.1
-α*x(2)*(1/((17.175/x(18))^β-1)) ≥ 0.2
α*x(4)*(1/((16.075/x(20))^β-1)) ≥ 0.1
α*x(2)*(1/((8.205/x(18))^β-1)) α*x(5)*(1/((17.197/x(21))^β-1)) ≥ 0.1
-α*x(3)*(1/((9.562/x(19))^β-1)) ≥ 0.2 α*x(6)*(1/((16.785/x(22))^β-1)) ≥ 0.1
α*x(7)*(1/((7.038/x(23))^β-1)) ≥ 0.1
α*x(6)*(1/((3.470/x(22))^β-1)) α*x(8)*(1/((16.794/x(24))^β-1)) ≥ 0.1
-α*x(3)*(1/((9.562/x(19))^β-1)) ≥ 0.2 α*x(9)*(1/((16.038/x(25))^β-1)) ≥ 0.1 (14)
α*x(10)*(1/((11.633/x(26))^β-1)) ≥ 0.1
α*x(13)*(1/((9.919/x(29))^β-1))
α*x(11)*(1/((19.904/x(27))^β-1)) ≥ 0.1
-α*x(4)*(1/((16.075/x(20))^β-1)) ≥ 0.2 α*x(12)*(1/((7.180/x(28))^β-1)) ≥ 0.1
α*x(16)*(1/((1.697/x(32))^β-1)) α*x(13)*(1/((18.288/x(29))^β-1)) ≥ 0.1
-α*x(4)*(1/((16.075/x(20))^β-1)) ≥ 0.2 α*x(14)*(1/((11.043/x(30))^β-1)) ≥ 0.1
α*x(15)*(1/((17.728/x(31))^β-1)) ≥ 0.1
α*x(2)*(1/((6.540/x(18))^β-1)) α*x(16)*(1/((9.734/x(32))^β-1)) ≥ 0.1
-α*x(5)*(1/((17.197/x(21))^β-1)) ≥ 0.2
C. User-defined Approach
α*x(4)*(1/((4.00/x(20))^β-1))
-α*x(5)*(1/((17.197/x(21))^β-1)) ≥ 0.2 In this study the desired range of TMS and PS are
considered as (0.1-1.1) and (0.5-2.0) [14]. Similarly, the
α*x(15)*(1/((6.188/x(31))^β-1)) operating range of α and β are taken as (0.14-80) and (0.02-
-α*x(6)*(1/((16.785/x(22))^β-1)) ≥ 0.2 (13) 2.00) respectively [5]. In this approach, the relay
characteristics coefficients (α, β) are also considered as
α*x(1)*(1/((4.970/x(17))^β-1)) decision variables along with TMS and PS. Therefore, each
-α*x(7)*(1/((7.038/x(23))^β-1)) ≥ 0.2 DOCR, consists of four decision variables. As a result, for 16
DOCRs, the total number of decision variables for a relay
α*x(10)*(1/((4.875/x(26))^β-1)) coordination problem using user-defined characteristics is 48.
-α*x(8)*(1/((16.794/x(24))^β-1)) ≥ 0.2 The objective function and relay constraint equations are not
included in this case due to page limitation. In this approach,
α*x(7)*(1/((3.896/x(23))^β-1))
the decision variables x(1)-x(16) are associated with TMS1-
-α*x(9)*(1/((16.038/x(25))^β-1)) ≥ 0.2 TMS16, x(17)-x(32) are associated with PS1-PS16, x(33)-
x(48) are associated with α1- α16, and x(49)-x(64) are
α*x(12)*(1/((5.214/x(28))^β-1))
associated with β1- β16.
-α*x(10)*(1/((11.633/x(26))^β-1)) ≥ 0.2
α*x(9)*(1/((9.296/x(25))^β-1))
V. RESULTS AND DISCUSSIONS
-α*x(11)*(1/((19.904/x(27))^β-1)) ≥ 0.2
In this section, the optimal relay settings are obtained via
α*x(14)*(1/((5.907/x(30))^β-1)) a linear and non-linear programming approach using the
-α*x(12)*(1/((7.180/x(28))^β-1)) ≥ 0.2 GWO algorithm, as shown in Table III. From the obtained
results, it can be seen that the optimal objective function (OF)
α*x(11)*(1/((11.996/x(27))^β-1)) values via a linear programming approach have been found
-α*x(13)*(1/((18.288/x(29))^β-1)) ≥ 0.2 to be 7.3169s, 2.5507s, and 1.7361s for NI, VI, and EI relay
characteristics, respectively. However, the optimal objective
α*x(3)*(1/((3.677/x(19))^β-1)) function values via a non-linear programming approach have
-α*x(14)*(1/((11.043/x(30))^β-1)) ≥ 0.2 been found to be 7.2453s, 2.5482s, and 1.7053s for the same
NI, VI, and EI relay characteristics, respectively. From the
α*x(16)*(1/((3.013/x(32))^β-1)) findings, it can be concluded that the non-linear programming
-α*x(14)*(1/((11.043/x(30))^β-1)) ≥ 0.2 approach gives better results as compared to the linear
programming approach for all kinds of standard relay

Authorized licensed use limited to: Motilal Nehru National Institute of Technology. Downloaded on September 07,2025 at 05:30:20 UTC from IEEE Xplore. Restrictions apply.
 characteristics. One more observation can be found that as we R2 0.6235 0.9285 47.8407 1.9511
change the relay characteristics (RC) from NI to VI and R3 0.1419 0.9427 63.8559 1.9459
R4 0.7675 0.8448 45.9244 1.9914
further EI, the total relay operating time is reduced. R5 0.2346 1.2716 49.4235 1.8263
Moreover, the optimal relay settings via a nonlinear R6 0.8736 0.6497 51.6842 1.8796
programming approach using user-defined relay R7 0.1131 1.4969 15.3540 1.8717
characteristics, as shown in Table IV, give the lowest value R8 0.1865 1.6036 42.5058 1.8651
of total primary relay operating time. From the obtained R9 0.5150 1.3554 22.7079 1.8950
R10 0.1017 1.8744 17.5383 1.6055
results, it can be concluded that as the number of decision
R11 0.3909 1.5497 45.8515 2.0000
variables increases, the optimization problem becomes more R12 0.1303 1.0340 55.7656 1.9918
flexible, and it gives the least optimal value of the objective R13 0.3007 1.9461 27.9315 1.9810
function of 1.6893s. The optimal primary and backup relay R14 1.0850 0.8164 15.2521 1.9480
operating times and measured coordination time (MCT) via a R15 0.1406 0.7856 59.7319 1.4227
linear programming approach using NI, VI, and EI relay R16 0.1022 1.0646 57.4066 1.8384
OF 1.6893s
characteristics are shown in figs. 2–4. The similar findings
obtained through a nonlinear programming approach are
shown in figs. 5–7. From the figs. 4 and 7, it can be seen that
the value of MCT for all the relay pairs are always greater
than 0.2 sec, which is one of the essential requirement of
optimal relay coordination between relay pairs. Similarly,
from figs. 2 and 5, it can be seen that primary relay operating
time for all the relay pairs are always greater than 0.1 sec,
which is also one of the essential condition of the optimal
relay coordination.

TABLE III. OPTIMAL OVERCURRENT RELAY SETTINGS USING GWO

LP NLP Fig. 2 Primary relay operating time in linear programming approach using
Relay

NI VI EI NI VI EI GWO
TMS TMS TMS TM PS TM PS TM PS
S S S
R 0.18 0.12 0.18 0.12 1.57 0.25 0.55 0.35 0.73
1 19 47 13 71 23 56 15 82 84
R 0.19 0.19 0.36 0.16 1.60 0.21 0.87 0.24 1.23
2 25 89 75 93 95 06 36 75 56
R 0.13 0.10 0.11 0.18 0.71 0.10 0.88 0.35 0.56
3 77 12 31 62 82 02 46 51 64
R 0.16 0.11 0.32 0.10 1.49 0.12 0.92 0.16 1.41
4 05 18 19 98 16 20 89 16 82
R 0.15 0.12 0.36 0.19 0.68 0.11 1.03 0.65 0.77
5 72 50 85 33 27 65 40 73 80
R 0.14 0.11 0.35 0.11 1.35 0.15 0.79 0.24 1.21 Fig. 3 Backup relay operating time in linear programming approach using
6 55 73 12 01 75 86 14 23 61 GWO
R 0.16 0.10 0.10 0.14 1.10 0.10 0.97 0.24 0.50
7 03 00 00 78 25 00 43 99 20
R 0.17 0.16 0.35 0.16 1.28 0.27 0.56 0.37 0.96
8 99 67 25 62 40 50 43 55 82
R 0.24 0.25 0.32 0.30 0.52 0.48 0.55 0.18 1.34
9 67 06 97 77 26 77 35 09 92
R 0.14 0.10 0.16 0.13 1.11 0.17 0.59 0.16 1.00
10 65 00 79 78 79 66 57 73 97
R 0.24 0.28 0.53 0.25 0.75 0.52 0.58 0.16 1.79
11 40 98 63 39 77 40 36 58 48
R 0.14 0.10 0.10 0.20 0.53 0.22 0.50 0.41 0.50
12 61 24 00 14 05 79 38 46 00
R 0.19 0.19 0.41 0.22 0.53 0.40 0.51 0.97 0.65
13 97 89 72 22 49 73 95 57 51 Fig. 4 MCT in linear programming approach using GWO
R 0.18 0.15 0.15 0.26 0.51 0.31 0.54 0.47 0.57
14 49 50 18 19 12 80 59 76 69
R 0.15 0.12 0.39 0.16 0.76 0.29 0.52 1.02 0.61
15 40 60 20 21 66 62 42 97 83
R 0.12 0.10 0.11 0.14 0.82 0.16 0.52 0.22 0.71
16 95 00 73 93 01 04 47 88 85
O 7.31 2.55 1.73 7.2453s 2.5482s 1.7053s
F 69s 07s 61s

TABLE IV. OPTIMAL OVERCURRENT RELAY SETTINGS USING GWO

Relay User-defined relay characteristics
TMS PS α β Fig. 5 Primary relay operating time in non-linear programming approach
R1 0.3450 1.2508 25.2130 1.9685 using GWO

Authorized licensed use limited to: Motilal Nehru National Institute of Technology. Downloaded on September 07,2025 at 05:30:20 UTC from IEEE Xplore. Restrictions apply.
 [1] R. Tiwari, R. K. Singh and N. K. Choudhary, "A Comparative Analysis
of Optimal Relay Coordination For Different Network
Configuration," 2021 1st International Conference on Power
Electronics and Energy (ICPEE), Bhubaneswar, India, 2021, pp. 1-6,
doi: 10.1109/ICPEE50452.2021.9358719.
[2] R. Tiwari, R. K. Singh and N. K. Choudhary, "A Comparative Analysis
of Optimal Coordination of Distance and Overcurrent Relays with
Standard Relay Characteristics using GA, GWO and WCA," 2020
IEEE Students Conference on Engineering & Systems (SCES),
Prayagraj, India, 2020, pp. 1-6, doi:
10.1109/SCES50439.2020.9236743.
Fig. 6 Backup relay operating time in non-linear programming approach [3] H. Sharifian, H. A. Abyaneh, S. K. Salman, R. Mohammadi and F.
using GWO Razavi, "Determination of the Minimum Break Point Set Using Expert
System and Genetic Algorithm," in IEEE Transactions on Power
Delivery, vol. 25, no. 3, pp. 1284-1295, July 2010, doi:
10.1109/TPWRD.2010.2043999.
[4] M. J. Damborg, R. Ramaswami, S. S. Venkata and J. M. Postforoosh,
"Computer Aided Transmission Protection System Design, Part I:
Algorithms," in IEEE Power Engineering Review, vol. PER-4, no. 1,
pp. 30-31, Jan. 1984, doi: 10.1109/MPER.1984.5525430.
[5] R. Tiwari, R. K. Singh, and N. K. Choudhary, “Optimal Relay
Coordination for DG-Based Power System Using Standard and User-
Defined Relay Characteristics”, Int. j. eng. technol. innov., vol. 12, no.
3, pp. 207-224, Mar. 2022.
[6] A. S. Noghabi, H. R. Mashhadi and J. Sadeh, "Optimal Coordination
of Directional Overcurrent Relays Considering Different Network
Fig. 7 MCT in Non-linear programming approach using GWO Topologies Using Interval Linear Programming," in IEEE
Transactions on Power Delivery, vol. 25, no. 3, pp. 1348-1354, July
2010, doi: 10.1109/TPWRD.2010.2041560.
VI. CONCLUSION [7] N. K. Choudhary, S. R. Mohanty, and R. Kumar Singh, “Coordination
of Overcurrent Relay in Distributed System for Different Network
In this paper, a comprehensive comparative analysis of Configuration,” J. Power Energy Eng., vol. 03, no. 10, pp. 1–9, 2015.
overcurrent relay coordination has been performed using [8] N. K. Choudhary, S. R. Mohanty, and R. K. Singh, “Impact of
distributed generator controllers on the coordination of overcurrent
linear and nonlinear programming approaches. From the relays in microgrid,” Turkish J. Electr. Eng. Comput. Sci., vol. 25, no.
obtained results, it can be seen that the nonlinear 4, pp. 2674–2685, 2017.
programming approach performs better as compared to linear [9] R. Tiwari, R. K. Singh and N. K. Choudhary, "Performance Analysis
programming. A tabular summary of the relay coordination of Optimization Technique for Protection Coordination in Single and
Multi-Loop Distribution System," 2019 International Conference on
study is given in Table V. From the results, it can be seen that Electrical, Electronics and Computer Engineering (UPCON), Aligarh,
in case II, a percentage reduction of 0.97 in NI, 0.098 in VI, India, 2019, pp. 1-6, doi: 10.1109/UPCON47278.2019.8980151.
and 1.77 in EI has been obtained while using GWO compared [10] S. Biswal, N. K. Sharma and S. R. Samantaray, "Optimal Overcurrent
to case I. The same findings were also achieved while using Relay Coordination Scheme for Microgrid," 2020 21st National Power
GA, as shown in Table IV. However, when comparison has Systems Conference (NPSC), Gandhinagar, India, 2020, pp. 1-6, doi:
10.1109/NPSC49263.2020.9331890.
been made between cases I and III, it can be seen that a
[11] M. B. Atsever, O. Karacasu and M. H. Hocaoglu, "Optimal Overcurrent
significant percentage reduction of 76.91 in NI, 33.77 in VI, Relay Coordination in Distribution Networks," 2021 56th International
and 2.69 in EI is achieved using the GWO algorithm. In the Universities Power Engineering Conference (UPEC), Middlesbrough,
context of future scope, the researchers can examine the relay United Kingdom, 2021, pp. 1-6, doi:
coordination process for different configurations of the test 10.1109/UPEC50034.2021.9548185.
system caused by line contingencies or islanded operation. [12] S. Roy, P. S. Babu and N. V. P. Babu, "Optimal Combined Overcurrent
and Distance Relays Coordination using Teaching Learning based
Optimization," 2017 14th IEEE India Council International
TABLE V. COMPARATIVE ANALYSIS OF RELAY COORDINATION Conference (INDICON), Roorkee, India, 2017, pp. 1-6, doi:
10.1109/INDICON.2017.8487876.
Programm
Cases RC GA GWO [13] R. Behkam, B. Vahidi, M. Zolfaghari, M. S. Naderi and G. B.
ing
NI 7.3504 7.3169 Gharehpetian, "HBBO-based Intelligent Setting and Coordination of
Directional Overcurrent Relays Considering Different
Case-I Linear VI 2.5608 2.5507 Characteristics," 2020 28th Iranian Conference on Electrical
EI 1.7400 1.7361 Engineering (ICEE), Tabriz, Iran, 2020, pp. 1-4, doi:
NI 7.2621 7.2453 10.1109/ICEE50131.2020.9260901.
Case-II Non-linear VI 2.5507 2.5482 [14] R. Tiwari, R. K. Singh and N. Kumar Choudhary, "Optimal
EI 1.7198 1.7053 Coordination of Dual Setting Directional Over Current Relays in
User- Microgrid With Different Standard Relay Characteristics," 2020 IEEE
Case-III Non-linear 1.6917 1.6893
defined 9th Power India International Conference (PIICON), Sonepat, India,
NI 1.20 0.97 2020, pp. 1-6, doi: 10.1109/PIICON49524.2020.9112883.
% reduction in case-II as
VI 0.39 0.098 [15] .Christie R. Power system test cases; 1993.
compared to case-I
EI 1.16 1.77 www.ee.washington.edu/research/pstca.
NI 76.98 76.91
% reduction in case-III [16] Mahamad Nabab Alam, Overcurrent protection of AC microgrids using
VI 33.94 33.77 mixed characteristic curves of relays, Computers & Electrical
as compared to case-I
EI 2.77 2.69 Engineering, Volume 74, 2019, Pages 74-88,
https://doi.org/10.1016/j.compeleceng.2019.01.003.
REFERENCES

Authorized licensed use limited to: Motilal Nehru National Institute of Technology. Downloaded on September 07,2025 at 05:30:20 UTC from IEEE Xplore. Restrictions apply.