Electrical Power and Energy Systems 78 (2016) 336–342

Contents lists available at ScienceDirect

Electrical Power and Energy Systems

journal homepage: www.elsevier.com/locate/ijepes

An application of a tent map initiated Chaotic Firefly algorithm

for optimal overcurrent relay coordination
S.S. Gokhale ⇑, V.S. Kale
Department of Electrical Engineering, Visvesvaraya National Institute of Technology (VNIT), Nagpur, India

a r t i c l e i n f o a b s t r a c t

Article history: Over-current relays provide primary as well as backup protection to electrical distribution systems. These
Received 2 December 2014 relays should be coordinated and set at the optimum values, to minimize the total operating time and
Received in revised form 19 November 2015 hence ensure that least damage is caused when a fault occurs. While they provide backup protection,
Accepted 25 November 2015
it is also imperative to ensure that their settings should not cause their inadvertent operation and sub-
Available online 21 December 2015
sequent sympathy trips.
This paper describes a Chaotic Firefly algorithm (CFA) for optimal time coordination of these relays. The
Keywords:
algorithm has been implemented in MATLAB and tested on several systems, out of which two have been
Power system protection
Over-current relay coordination
illustrated in this paper. The results obtained by the Chaotic Firefly algorithm are compared with those
Optimization obtained by the conventional Firefly algorithm (FA). The novel feature of this paper is the application
Firefly algorithm of the Chaotic Firefly algorithm to the problem of over-current relay coordination.
Chaos theory Ó 2015 Elsevier Ltd. All rights reserved.
Tent maps

Introduction optimal coordination of relays is formulated as a relay operating

Electric power distribution systems are provided primary and
back up protection by over current relays. Sometimes they are also
used for providing back-up protection to transmission systems,
which employ distance protection as the primary protection. These
relays while meeting the requirements of sensitivity, selectivity
and reliability have to operate fast enough without mal-
operating, so that the faulty section is isolated immediately by
the circuit breaker. This makes coordination of these relays impor-
tant [1].
The relay coordination problem is often stated as:
If the magnitude of the load and fault current at every bus are
known, how to set the relays at the different buses, so that
the entire system gets primary as well as backup over-current
protection?
Correct coordination is necessary to ensure that the relays do
not operate out of turn. Optimal coordination is necessary so that
the system has the lowest possible fault clearing time. As the
size and complexity of the system goes on increasing it becomes
more and more difficult to coordinate the relays. Therefore, the
⇑ Corresponding author. Tel.: +91 712 2241175; fax: +91 7104 232376.
E-mail addresses: sanjeev_gokhale@rediffmail.com (S.S. Gokhale), tovijay_kale@
rediffmail.com (V.S. Kale).
time minimization problem under selectivity and other constraints
[2,3].
A good review of time-overcurrent relay coordination has been
presented in [4]. Usage of digital relays has made possible the set-
ting of the Time Multiplier Setting (TMS) at any non-integral value.
This problem has been dealt with by several optimization meth-
ods in the past. In [5] a linear programming technique for adaptive
protection has been applied. A random search technique has been
employed in [6]. With the advent of meta-heuristic algorithms,
researchers began applying these to the relay coordination prob-
lem. Evolutionary algorithms were the first to appear [7]. Genetic
algorithm (GA) were applied successfully for relay coordination
[8–13]. The Particle Swarm Optimization algorithm (PSO) was
applied in [14]. The Artificial Bee colony algorithm by [15], a hybrid
evolutionary algorithm based on Tabu search by [16], and a Seeker
algorithm [17] were used to solve the relay coordination problem.
Recently the authors applied the Firefly algorithm (FA) for the
same and compared it with Genetic Algorithms and Linear Pro-
gramming [18].
This paper proposes the Chaotic Firefly algorithm using tent
maps for optimally coordinating the relays. The chaos theory is
well established and dealt in with details in [19]. In FA, the quality
of solutions mainly depends on the randomization parameters a
and attractiveness b between the fireflies. If these values are not
properly selected it leads to premature convergence. To improve
the quality of solutions, in this paper, chaos theory is incorporated

http://dx.doi.org/10.1016/j.ijepes.2015.11.087
0142-0615/Ó 2015 Elsevier Ltd. All rights reserved.
 S.S. Gokhale, V.S. Kale / Electrical Power and Energy Systems 78 (2016) 336–342
to prevent the search process from being trapped in local minima.
Chaos theory is introduced by modifying the concept of random
movement factor variable a.
Two relay coordination cases have been investigated by apply-
ing the conventional Firefly algorithm as well as the Chaotic Firefly
algorithm on them. Detailed calculations have been shown.
Principle of over-current relay coordination
As an example, in the two section radial system shown in Fig. 1,
for a fault beyond bus P (zone k), the relay RP at bus P, operates
first. The relay RQ at bus Q, serves the purpose of back-up protec-
tion. As an example if the operating time of relay RP, is set at
0.1 s, the relay RQ is set such that it should operate at 0.1 s. Plus
the operating time of circuit breaker at bus P, plus the overshoot
time of relay RQ. Overshoot occurs only for analog relays and not
for digital relays.
In general the coordination of directional over-current relays in
power systems can be stated as an optimization problem, where
the sum of operating times of the relays ‘z’ of the system, for a fault
is to be minimized. Generally near end faults are considered.
Mathematically it can be stated as:
Minimize :
Xm
z¼ W p  tp;k :
p¼1
Here m denotes the total number of relays; z is the objective func-
tion to be minimized; tp,k is the operating time of the primary relay,
for fault in zone k, for near end fault, and Wp is the weight assigned
for operating time of the relay Rp.
As in distribution systems the lines are short and are nearly
equal in length, the probability of occurrence of the fault is equal
for all the lines and hence equal weights are assumed and W = 1
is assigned to all the relays [5,20]. This is the general approach.
We have also to take into consideration the constraints in the
problem:
A. Coordination constraints
The relays in the system have to be coordinated by the criterion,
t q;k  tp;k P STI;
where tp,k is the operating time of the primary relay at k, for
near end fault and tq,k is the operating time of the back-up
relay, for the same fault. STI is the Selective Time Interval
and is taken between 0.1 and 0.5 s. This depends on the speed
with which the circuit breakers operate.
B. Bounds on the relay operating time
Relays take a certain minimum time to operate. However, they
should not take too long a time to operate. This puts the
constraint,
t p;k;min 6 t p;k 6 tp;k;max ;
Fig. 1. A radial distribution system.
337
where tp,k,min is the minimum operating time and tp,k,max is
the maximum operating time of relay at bus P for near end
fault. Similarly, it is also for other relays.
C. Relay characteristics
The relays are assumed identical, having normal Inverse Def-
inite Minimum Time (IDMT) characteristics, which can be
described by the equation,
0:14 ðTMSÞ
t op ¼ ; ð4Þ
PSMð0:02Þ  1
where top is the relay operating time. The constants 0.14 and
0.02 are standard for Inverse Definite Minimum Time (IDMT)
relay characteristics. TMS is the time multiplier setting and
PSM is the plug setting multiplier.
Irelay
PSM ¼ : ð5Þ
PS
In the above form of Eq. (4), the problem is non-linear in nat-
ure. The relay coordination problem can be formulated as a
Linear programming problem by considering the Plug Setting
(PS) of the relays fixed, and the operating time of the relays,
a linear function of the Time Multiplier Setting. (This simplifi-
cation need not be done always and the problem can be for-
mulated and solved as a non-linear problem [11].)
In the linear problem as PSM is fixed, (4) reduces to
ð1Þ
top ¼ ap ðTMSÞ; ð6Þ
where
0:14
ap ¼ : ð7Þ
PSMð0:02Þ  1
Hence, the objective function can be written as,
Minimize :
Xm
z¼ ap ðTMSÞp ; ð8Þ
p¼1
subject to the constraints imposed by (2) and (3).
Coordination of relays in ring fed systems
ð2Þ A typical single end fed ring main feeder used in distribution
systems is shown in Fig. 2. It has the advantage that continuity
of supply is maintained to all loads even if a fault occurs on any
bus. In such arrangements, some of the relays also require a direc-
tional feature to implement the protective zones. Relays 2, 4, 6, 8
form one group of relays to be coordinated while 1, 3, 5, 7 form
another group. These loops had to be separately coordinated previ-
ously but with the use of optimization algorithms, there is no such
need [4,21].
Chaos theory and the Firefly algorithm
ð3Þ
Firefly algorithm (FA)
The Firefly algorithm (FA) is a meta-heuristic algorithm formu-
lated by Yang in 2007. It is a swarm intelligence type of algorithm
inspired by the flashing behavior of fireflies [22].
FA initially produces a swarm of fireflies located randomly in
the search space. The position of each firefly in the search space
represents a potential solution of the optimization problem. The
dimensions of the search space are equal to the number of optimiz-
ing parameters. The objective function depends upon the position
of the firefly. As positions change every iteration, the objective
value also changes. The intensity of brightness of each firefly
depends on the objective value of that firefly.
338 S.S. Gokhale, V.S. Kale / Electrical Power and Energy Systems 78 (2016) 336–342
Fig. 2. A ring main feeder.
Each fly is attracted by the brightness of the other flies and
moves to the fly that relatively ‘appears’ brighter. This leads to
the concept of attractiveness. Attractiveness depends on brightness
as well as distance. A brighter firefly far off may not be as attractive
as a less bright firefly that is near. The velocity with which a firefly
moves towards another firefly depends on the attractiveness (b).
Each iteration the algorithm computes the intensity of brightness
and attractiveness of each firefly. Depending on these values, the
positions of the fireflies are updated. After sufficient iterations,
all fireflies converge to the best possible position in the search
space.
A firefly is specified by
xi ¼ ðxi1 ; . . . ; xiD Þ for i ¼ 1; . . . ; N;
where N denotes the population size and D is the dimensionality of
the problem [23].
The initial population of fireflies is generated from a uniform
distribution by the equation:
ð0Þ
xij ¼ Uð0; 1Þ  ðUbj  lbj Þ þ lbj for i ¼ 1; . . . ; N;
where U(0, 1) denotes a random number drawn from a uniform dis-
tribution in the interval [0, 1], lbj and Ubj are the lower and upper
limits of the corresponding jth problem variable.
As stated, the intensity of brightness of the mth firefly Im is
related to the objective function. The relationship is inversely pro-
portional for minimization problems
1
Im ¼ f fFitnessðX m Þg :
In this paper, FA searches for an optimal solution by maximizing
the function ‘Light Intensity’ (as FA originally dealt with maximiza-
tion problems) that is obtained from the relay operating time min-
imization function by:
Maximize :
1
Im ¼ E:
1þz
The attractiveness between the mth and nth firefly, bm,n is given
by,
bm;n ¼ ðbmax;m;n  bmin;m;n Þ expðcr2m;n Þ þ bmin;m;n ; ð13Þ
where rm,n is the Cartesian distance between the mth and nth fly.
If the light intensity of nth firefly is larger than the intensity of
the mth firefly, then the mth firefly moves towards the nth firefly
and its position at the kth iteration is denoted by the following
equation:
xm ðkÞ ¼ xm ðk  1Þ þ bm;n ðxm ðk  1Þ  xn ðk  1ÞÞ þ aðrand  0:5Þ;
ð14Þ
where a is the random movement factor. A large a explores the dis-
tant search space, whereas a smaller a facilitates local search. b is
limited between 0 and 1. If bmax is kept at 1, the algorithm performs
cooperative local search with the brightest firefly strongly deter-
mining the position of the other fireflies. If bmax is set to zero, there
is only non-cooperative random search.
The absorption coefficient c determines the variation of attrac-
tiveness with distance. Setting c to 0 corresponds to no variation or
attractiveness is constant. Putting c = 1 results in attractiveness
equal to zero which means, a complete random search. In this
paper, c is chosen as 1. Choice of these parameters is important
for the fast convergence of the algorithm. Eqs. (13) and (14) pro-
duces better offspring every iteration. The penalty method is used
for handling the constraints [24].
Chaos theory
Chaos theory studies the behavior of dynamical systems that
are highly sensitive to initial conditions. Small differences in initial
conditions like those due to rounding errors in numerical compu-
tation yield widely diverging outcomes for such systems. Chaos
is non-periodic, non-converging and bounded. Chaos is more struc-
tured than random phenomena, which are very haphazard. Chaos
works better than noise [25]. Chaos has been introduced in opti-
mization and yields robust global optimization [26]. In electrical
engineering chaos has been used in load forecasting, economic load
dispatch, parameter identification of transformers [27–29]. Chaos
can be generated using tent maps, logistic maps etc. In this paper,
the tent map function is used to generate chaos. The expression of
ð9Þ the tent map is given by:

anþ1 ¼ 2lan for an < 0:5
; ð15Þ
anþ1 ¼ 2lð1  an Þ for an P 0:5
where ‘n’ denotes the iteration number and l is a positive real con-
stant. The value of an are generated under the condition that the ini-
ð10Þ tial an e (0, 1). A tent map for l = 0.88 is shown in Fig. 3. As the
iterations, progress alpha takes new values.
ð11Þ
ð12Þ
Fig. 3. A tent map.
 S.S. Gokhale, V.S. Kale / Electrical Power and Energy Systems 78 (2016) 336–342 339

The difference between the conventional Firefly algorithm and

the Chaotic Firefly algorithm lies in the way the initial population
is generated. The conventional Firefly algorithm uses a random
number generator to generate the population, whereas in the
Chaotic Firefly algorithm, the user initializes the initial alpha
(a0). In Eq. (10) instead of using random numbers, we use chaotic
distribution and the subsequent alphas (a) are generated by the
tent map function. The final solutions of the relay coordination
problem are certain to lie within an area bounded by the various
constraints. Therefore, they are deterministic and so it is better
to approach them using chaotic number generation than a random
approach which is employed by the conventional Firefly algorithm.
However, the pivotal concept of generating populations of solu-
tions of evolutionary algorithms is retained.
The structure of the Chaotic Firefly algorithm is shown below:

Fig. 4. A parallel feeder system (Case 1).

Define Objective function f(x);
Create Initial population of fireflies; Initialize a0.
Formulate light intensity Im, associated with f(x) Table 1
Define absorption coefficient c Currents seen by the relays and the aP constants for Case 1.
While (t < Maximum _Iterations)
Fault point Relay
Update a using the tent map formula
1 2 3 4 5
for i = 1:n (all n fireflies)
for j = 1:n (n fireflies) A Irelay 9.059 3.019 – 3.019 –
aP 3.106 6.265 – 6.265 –
if (Ij > Ii),
move firefly i towards j; B Irelay 3.019 – 9.059 3.019 –
aP 6.265 – 3.106 6.265 –
end if
Vary attractiveness with distance r via; exp(cr2) C Irelay 4.875 – 4.875 4.875 29.25
aP 4.34 – 4.34 4.34 2.004
Evaluate new solutions and update light intensity;
end for j ‘–’ indicates fault is not seen by the relay.
end for i
Rank fireflies and find the current best;
end while Coordination constraints
end 6:265x4  3:106x2 P 0:2; ð17Þ

6:265x1  3:106x3 P 0:2; ð18Þ

Results and discussions

In order to study the Chaotic Firefly algorithm as well as the

4:341x1  2:004x5 P 0:2; ð19Þ
Firefly algorithm two cases have been taken here. Detailed data
about the cases can be found in Ref. [9]. 4:341x4  2:004x5 P 0:2: ð20Þ
The Bounds on the relay operating time are
Case 1
3:106x1 P 0:1; ð21Þ
The case of the parallel feeders, fed from a single end is tested
by the Firefly algorithm as well as the Chaotic Firefly algorithm. 6:265x2 P 0:1; ð22Þ
Fig. 4 shows this system. Relays 1, 4 and 5 are non-directional
relays, whereas 2 and 3 are directional relays. The relays are 3:106x3 P 0:1; ð23Þ
assumed to have plug setting of 1 and the corresponding current
transformer ratios are 300:1. A and B are two fault locations in 6:265x4 P 0:1; ð24Þ
the middle of the lines shown. Relay 4 will back up relay 2 for fault
at A and relay 1 will back up relay 3 for fault at B. Total fault 2:004x5 P 0:1: ð25Þ
current in each case is assumed to be 4000 A. The current This dimensionality of the problem is four. For solving the case
divides amongst the two parallel paths (3000 A and 1000 A.) in both the FA and CFA were applied. In each case, the population size
proportion to their impedances. The STI is assumed as 0.2 for both was taken as 40 [30] and the number of iterations was taken to be 50.
the cases. The equations were solved using the Firefly algorithm with a taken
Three possible fault locations A, B, C are considered and the cor- at a constant value of 0.5 and bmin = 0.2 and c = 1. The convergence
responding relay currents and ap constants of the relays are shown characteristic for ‘z’ when FA was applied is shown in Fig. 5.
in Table 1. The relay TMS’s are denoted through x1–x5. When the same equations were solved using the CFA, a was ini-
By choosing the appropriate ‘aP’ constants from Table 1. The tialized at a value of 0.95. Thereafter a takes values as decided by
objective function is: the tent map function. Fig. 6 shows the variation of a as the itera-
Minimize : tions progress and the convergence characteristic for ‘z’ when CFA
is applied. As the iterations are integer values ‘z’ are not shown
z ¼ 3:106x1 þ 6:265x2 þ 3:106x3 þ 6:265x4 þ 2:004x5 ; ð16Þ
continuous, but alpha has been shown interpolated in the graph
subject to the: to illustrate their tent map nature.
340 S.S. Gokhale, V.S. Kale / Electrical Power and Energy Systems 78 (2016) 336–342

because depending on the location of the fault different primary

Fig. 5. Convergence characteristics with FA for Case 1.
backup relay configurations are formed. These depend on the
direction of the current in the various feeders. The dimensionality
of the problem is six.
The line data for the system is given in Table 3. The line charg-
ing admittances are neglected. For case, four different fault loca-
tions are taken. The primary back up relations between the
relays for the faults are given in Table 4, the CT ratios and the Plug
settings are shown in Table 5. The currents seen by the relays and
the ap constants for the four different fault locations A, B, C, D are
shown in Table 6. A, B, C, D are assumed in the middle of the
respective feeders. Only one back up relay is considered for every
fault. Only relay 1 and 5 have a directional feature.
This case has eleven constraints: six constraints due to bounds
on relay operating time and five arise because of the coordination
condition. Minimum time of operation of every relay is taken as
0.1 s. The range of TMS is between 0.025 and 1.2 s. For this case,
the selective time interval is taken as 0.3 s. The relay TMS are
denoted through x1–x6.
For this case the minimization problem becomes:
Minimize :
z ¼ 102:4x1 þ 6:06x2 þ 98:75x3 þ 24:4x4 þ 35:31x5 þ 11:53x6 :ð26Þ
The constraints arising because of minimum operating time of
relay are:
3:646x1 P 0:1; ð27Þ

6:055x2 P 0:1; ð28Þ

8:844x3 P 0:1; ð29Þ

8:844x4 P 0:1; ð30Þ

4:044x5 P 0:1; ð31Þ

11:539x6 P 0:1; ð32Þ

Fig. 6. Convergence characteristics and variations in alpha with CFA for Case 1.

Table 2
Comparison for TMS and total operating time for the relay system considered in Case
1 obtained using FA and CFA.

Time multiplier settings Firefly algorithm Chaotic Firefly algorithm

TMS 1 0.032 0.032
TMS 2 0.016 0.047
TMS 3 0.121 0.091
TMS 4 0.016 0.016
TMS 5 0.104 0.094
Total operating time ‘z’ (s) 1.73 1.63
P
Net gain D(t)s 0.1 s

As clearly seen with CFA the convergence is faster and yields a

better solution for ‘z’.
The TMS obtained for the relays and the total operating time of
all the relays given by (1) is shown for the two algorithms in
Table 2.
As shown, the Chaotic Firefly algorithm gives an advantage of
0.1 s over the Firefly algorithm. This may appear meager, but is sig-
nificant considering that it is a very small system.

Case 2

Next, a single end fed distribution system with six over-current

relays is considered (Fig. 7). The system is a multi loop system Fig. 7. A multi-loop distribution system (Case 2).
 S.S. Gokhale, V.S. Kale / Electrical Power and Energy Systems 78 (2016) 336–342 341

Table 3
Line data for Case 2.

Line Impedance (X)

1–2 0.08 + jl
2–3 0.08 + jl
1–3 0.16 + j2

Table 4
Primary–backup relationships of relays in Case 2.

Fault point Primary relay Backup relay

A 1 –
2 4
B 3 1
4 5
C 5 –
6 3
D 3 1 Fig. 8. Convergence characteristics with FA for Case 2.
5 –

Table 5
CT ratios and plug settings of relays for Case 2.

Relay CT ratio (A/A) Plug setting

1 1000/1 1
2 300/1 1
3 1000/1 1
4 600/1 1
5 600/1 1
6 600/1 1

Table 6
Relay currents and aP constants for Case 2.

Fault point Relay

1 2 3 4 5 6
A Irelay 6.579 3.13 – 1.565 1.565 –
aP 3.646 6.065 – 15.55 15.55 –
B Irelay 2.193 – 2.193 2.193 2.193 –
aP 8.844 – 8.844 8.844 8.844 – Fig. 9. Convergence characteristics and variation in alpha with CFA for Case 2
C Irelay 1.096 – 1.096 – 5.482 1.827 (alpha initialized to 0.95).
aP 75.91 – 75.91 – 4.044 11.53
D Irelay 1.644 – 1.644 – 2.741 –
Table 7
aP 13.99 – 13.99 – 6.872 –
Comparison for TMS and total operating time for the relay system considered in Case
‘–’ indicates fault is not seen by the relay. 2 using FA and CFA.

Time multiplier settings Firefly algorithm Chaotic Firefly algorithm

Solutions of the constraints from (28)–(32) violate the con-
straints of minimum value of TMS (All TMS’s should be >0.025). TMS 1 0.027 0.027
TMS 2 0.130 0.221
Therefore, they are upward revised to 0.025. These constraints
TMS 3 0.025 0.025
are rewritten as: TMS 4 0.025 0.025
TMS 5 0.025 0.029
x2 ; x3 ; x4 ; x5 ; x6 P 0:025; ð33Þ TMS 6 0.489 0.363
The coordination constraints are: Total operating time ‘z’ (s) 16.25 14.69
P
Net gain D(t)s 1.56 s
15:55x4  6:065x2 P 0:3; ð34Þ
tions was taken to be 60. The equations were solved using the Fire-
8:844x1  8:844x3 P 0:3; ð35Þ
fly algorithm with a taken at a constant value of 0.9 and bmin = 0.2
and c = 1. The convergence characteristic for ‘z’ when FA is applied
8:844x5  8:844x4 P 0:3; ð36Þ
is shown in Fig. 8.
When the same equations are solved using the CFA, a is initial-
75:91x3  11:53x6 P 0:3; ð37Þ
ized at a value of 0.95. Thereafter a takes value as decided by the
tent map function. Fig. 9 shows the convergence characteristic
13:998x3  13:998x1 P 0:3: ð38Þ
for ‘z’ and variation of a as the iterations progress for CFA.
The problem was solved using the conventional Firefly algo- As clearly seen with CFA the convergence is faster and in less
rithm (FA) and the Chaotic Firefly algorithm (CFA). In each case, iterations yields a better solution for ‘z’. The results obtained for
the number of fireflies was taken as 40 and the number of itera- the TMS and total operating time are shown in Table 7.
342 S.S. Gokhale, V.S. Kale / Electrical Power and Energy Systems 78 (2016) 336–342
Fig. 10. Convergence characteristics and variations in alpha with CFA for Case 2
(alpha initialized to 0.15).
The Chaotic Firefly algorithm gives an advantage of 1.56 s over
the plain Firefly algorithm.
Effect of varying a0 on the solution
The results were tested for different initializations of a0. As a
sample, a graph of a and the corresponding convergence character-
istics with a0 initialized at 0.15 are shown for case 2 in Fig. 10. It
was found that the best results were obtained when a0 was closer
to 1 than to zero.
Conclusion
The Firefly algorithm and the Chaotic Firefly algorithm have
been applied to solve the relay coordination problem. The pro-
grams have been developed in MATLAB, to solve the problem for
various relay configurations. Two sample cases have been illus-
trated in this paper. As the number of relays goes on increasing,
the highly constrained nature of the problem becomes more evi-
dent. Also a more exact optimal coordination minimizes the total
operating time. This is of prime importance because, as this time
decreases, the damage caused by the fault can get limited. In addi-
tion, unwarranted tripping of the circuit breakers can be avoided.
As can be seen from the figures and tables, the Chaotic Firefly
algorithm yields better results than the conventional Firefly algo-
rithm on the points of quality of solution, speed of convergence
and number of iterations taken to obtain the best solution. Further-
more, the results obtained show that the Chaotic Firefly algorithm
is superior to the conventional Firefly algorithm irrespective of the
initialization value of a0.
This work justifies the approach that it is better to use chaos
than random initialization while using evolutionary algorithms
for relay coordination.
References
[1] Blackburn J. Protective relaying, principles and applications. New York: Marcel
Dekker Inc; 1987.
[2] Urdaneta AJ, Restrepo H, Marquez S, Sanchez J. Optimal coordination of
directional relays in interconnected power systems. IEEE Trans Power Delivery
1988;3(3):903–11. July.
[3] Perez Luis G, Urdaneta AJ. Optimal coordination of Directional overcurrent
relays considering definite time back-up relaying. IEEE Trans Power Delivery
1995;14(4):1276–84.
[4] Birla D, Maheshwari RP, Gupta HO. Time overcurrent relay coordination: a
review. Int J Emerg Electr Power Syst 2005;2(2):1–13 Article 1039.
[5] Chattopadhyay Bijoy, Sachdev MS, Sidhu TS. An online relay coordination
algorithm for adaptive protection using linear programming technique. IEEE
Trans Power Delivery 1996;11(1):165–73.
[6] Birla D, Maheshwari RP, Gupta HO, Deep K, Thakur M. Applications of random
search technique in directional overcurrent relay coordination. Int J Emerg
Electr Power Syst 2006;7(1) (Article 1).
[7] So CW, Li KK. Time coordination method for power system protection by
evolutionary algorithm. IEEE Trans Ind Appl 2000;36(5):1235–40.
[8] Koochaki A, Asadi MR, Naghizadeh RR. Optimal overcurrent relay coordination
using Genetic Algorithm. In: The proceedings of the 11th international
conference OPTIM; 2008. p. 197–202.
[9] Bedekar PP, Bhide SR. Optimum coordination of over-current relays using the
continuous genetic algorithm. Expert Syst. Appl. 2011;38:11286–92 (Elsevier).
[10] So CW, Li KK, Lai KT, Fung KY. Applications of genetic algorithm to over-
current relay grading coordination. In: The proceedings of the 4th
international conference on advances in power systems, control, operation
and management, Hong Kong; 1997. p. 283–7.
[11] Bedekar PP, Bhide SR. Optimal coordination of directional overcurrent relay
using GA-NLP hybrid approach. IEEE Trans Power Delivery 2011;26
(1):109–19.
[12] Razavi Farzad, Abyaneh Hossein Askarian, Al-Dabbagh Majid, Mohammadi
Reza, Torkaman Hossein. A new comprehensive genetic algorithm method for
optimal overcurrent relays coordination. Electr Power Syst Res 2008;78:
713–20.
[13] Abbas Saberi Noghabi, Javed Sadeh, Habib Rajabi Mashhadi. Considering
different network topologies in optimal over-current relay coordination using
a hybrid GA. IEEE Trans Power Delivery 2009;24(4):1857–63.
[14] Liu An, Yang Ming-Ta. Optimal coordination of directional over-current relay
using NM-PSO technique. In: IEEE transactions on international symposium on
computer and control; 2012. p. 678–81.
[15] Uthitsunthorn D, Pao-L-Or P, Kulworawanichpong T. Optimal over-current
relay coordination using artificial bees colony algorithm. In: The proceedings
of the eighth electrical engineering, electronics, computer, telecommunications
and information technology conference; 2011. p. 901–4.
[16] Xu Chunlin, Zou Xiufen, Wu Chuansheng. Optimal coordination of protection
relay using new hybrid evolutionary algorithm. In: IEEE transactions on CEC;
2008. p. 823–8.
[17] Turaj A. Coordination of directional over-current relay using seeker algorithm.
IEEE Trans Power Delivery 2012:1415–22.
[18] Gokhale SS, Kale VS. An application of the Firefly algorithm for optimal
overcurrent relay coordination. In: The proceedings of OPTIM 2014
conference, Bran, Romania, May 2014. p. 150–154. http://dx.doi.org/10.1109/
OPTIM.2014.6850887.
[19] Schuster HG. Deterministic chaos: an introduction. 2nd revised ed. Physick-
Verlag Gmnh; 1988. D 6940 Weinheim, Federal Republic of Germany.
[20] Zeineldin HH, El- Saadany EF, Salama MMA. Optimal coordination of
overcurrent relays using a modified particle swarm optimization. Electr
Power Syst Res 2006;76:988–95.
[21] Soman SA. Lectures on power system protection, Module 5, lecture 19, NPTEL
Online. <www.cdeep.iitb.ac.in/> [10.01.2010].
[22] Yang XS. Nature inspired meta-heuristic algorithms. 2nd ed. Beckington:
Luniver Press; 2010.
[23] Fister Jr Iztok, Perc Matjaz, Kamal Salahuddin M. A review of chaos based
Firefly algorithms: perspectives and research challenges. Appl Math Comput
2015;252:155–65 (ISSN 0096-3003).
[24] Rao SS. Engineering optimization: theory and practice. third ed. India: New
Age International Pvt. Ltd; 1998.
[25] Bucolo Maide, Caponetto Riccardo, Fortuna Luigi, Frasca Mattia, Rizzo
Alessandro. Does chaos work better than noise. IEEE Circuits Syst Mag
2002;2(3):4–19. http://dx.doi.org/10.1109/MCAS 2002 1167624.
[26] Gandomi AH, Yang XS, Talathari S, Alavi AH. Firefly algorithm with chaos.
Commun Non-linear Sci Numer Simulat 2013;18(1):89–98.
[27] Guio-Ching-Liao. Application of a Fuzzy Neural network combined with a
chaotic Genetic algorithm and simulated annealing to short term load
forecasting. IEEE Trans Evol Comput 2006;10(3):330–40.
[28] Arul R, Velusami S, Ravi G. Chaotic Firefly algorithm to solve economic load
dispatch problems. In: IEEE international conference on green computing,
communications and conservation of energy; 2013. p. 458–64. http://dx.doi.
org/10.978-1-4673-6126-2/13.
[29] Rashtchi V, Rahimpour E, Fotoohabadi H. Parameter identification of
transformer detailed model based on chaos optimization algorithm. IET
Electr Power Appl 2011;5(2):238–46. http://dx.doi.org/10.1049/iet-epa.2010.
0147.
[30] Yang Xin-She. Cuckoo search and Firefly algorithm: theory and applications.
London: Springer; 2014.