Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

This article has been accepted for publication in a future issue of this journal, but has not been

fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2018.2821175, IEEE
Transactions on Industrial Informatics
1

Dual-Setting Directional Overcurrent Relays for


Protecting Automated Distribution Networks
Amin Yazdaninejadi, Sajjad Golshannavaz, Daryoush Nazarpour, Saeed Teimourzadeh, Student
Member, IEEE, Farrokh Aminifar, Senior Member, IEEE

 conditions [6]. Deterministic method has also been deployed for


Abstract-- This paper elaborates a new protection scheme the same mission which can provide reliable and accurate
based on dual-setting DOCRs for automated distribution results [7]. However, such a method is in challenge with
networks. In this course, the reduction in relays operation time increased complexity; thus, it is not an efficient choice for real-
saturates as the number of these relays increases. Thus, optimal
scale meshed distribution networks. Some recent studies have
deployment of dual-setting DOCRs should be specified in an
efficient manner. To do so, a multi-objective optimization also contributed by developing optimization frameworks [8],
approach is established which compromises the reduction of total [9]. These methods have made progress in devising efficient
operation time and the number of dual-setting DOCRs. protection strategies; however, they could also be further
Coordination constraints are accommodated and non-standard improved. The research conducted in [10] has classified the
inverse-time characteristics are established to intensify flexibility protection coordination violations in threefold. Then, a multi-
of the proposed strategy. The proposed model lies within a non-
point strategy was developed to alleviate the number of
linear programming fashion which is tackled by particle swarm
optimization (PSO). Moreover, the augmented ε-constraint violations. As another solution, heuristic algorithms have been
approach is utilized to reach the Pareto-optimal solutions proposed for optimal coordination of conventional DOCRs [6],
following which the fuzzy decision making process determines the [11]-[13]. In [14], it is shown that effective tuning of heuristic
best compromised solution. Detailed simulation studies are algorithms results in more suitable solutions. Some hybrid
carried out to interrogate performance of the proposed algorithms such as hybrid particle swarm optimization–
approach.
gravitational search algorithm and gravitational search
algorithm-sequential quadratic programming are further
Index Terms—Dual-setting directional over current relay,
optimal deployment, non-standard characteristics. explored in [15], [16]. An approach was presented in [17] which
splits the distribution network to smaller clusters to speed up the
I. INTRODUCTION coordination process. However, this technique could deteriorate
accuracy of the results against the centralized manner. What’s
So far, directional overcurrent relays (DOCRs) are
more, new generation of relay are able to provide non-standard
effectively deployed for reliable protection of meshed
time-current characteristics which adds to flexibility of the
distribution systems. Modern distribution systems also call for
coordination process [18], [19]. Application of non-standard
directional protections due to vast penetration of distributed
coordination strategy for conventional DOCRs was investigated
generations (DGs) [1], [2] rendering bidirectional power flows
in [9], [20]. The obtained solution is a non-standard inverse-
[3]. Furthermore, integration of DG units increases short-circuit
time characteristic with enhanced protection quality.
level which in turn, would damage the devices more intensely.
Despite improvements, attaining a faster protection task is
Accordingly, faster protection schemes are of a crucial
still of high concern. A fast, selective, and reliable protection
significance to avert such conditions [4], [5].
system reduces the level of power equipment damage, averts the
To date, numerous studies have been conducted in the area
unintentional feeder or DG disconnections, and further
of conventional DOCRs coordination. In this context, trial and
improves the reliability metrics [21]. Recent technological
error methods have been thoroughly examined. Although these
advances in automation and communication infrastructure have
approaches have resulted in an enhanced protection scheme,
paved the way for fast-response protection schemes [22]. A
they are encountered with serious problems in changing
communication-assisted overcurrent protection scheme has been
presented in [23] for radial distribution systems with distributed
Manuscript received August 19, 2017; revised January 20, 2018, February 20, generation. In [24], an approach is presented to provide an
2018; accepted March 11, 2018. Copyright © 2018 IEEE. Personal use of this accurate and low-maintenance solution for SCADA-based
material is permitted. However, permission to use this material for any other
purposes must be obtained from the IEEE by sending a request to pubs-
protection task. Recently, dual-setting DOCRs are
permissions@ieee.org. manufactured as competent alternatives to further improve the
A. Yazdaninejadi, S. Golshannavaz and D. Nazarpour are with the Electrical capabilities of conventional DOCRs [25], [26].
Engineering Department, Urmia University, Urmia 5756151818, Iran (e-mail:
Against the conventional DOCRs, a single unit dual-setting
a.yazdaninejadi@urmia.ac.ir, s.golshannavaz@urmia.ac.ir d.nazarpour@urmia.ac.ir).
S. Teimourzadeh and F. Aminifar are with the School of Electrical and Computer DOCR offers both the primary and backup functionalities,
Engineering, College of Engineering, University of Tehran, Tehran, 11365-4563, Iran concurrently. Thus, each dual-setting DOCR is programmed in
(e-mail: teimourzadeh@ut.ac.ir, faminifar@ut.ac.ir). Correspondence to: D. two directions namely forward/primary and reverse/backup,
Nazarpour, Electrical Engineering Department, Urmia University, Urmia, Iran (e-
mail: d.nazarpour@urmia.ac.ir). abbreviated by “fw/p” and “rv/b”, respectively. Each direction

1551-3203 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2018.2821175, IEEE
Transactions on Industrial Informatics
2

works independently; hence, some of the interlinked operation time of relays with a lower deployment level of
coordination constraints are relaxed. This feature contributes to dual-setting DOCRs to reduce the level of power equipment
an improved performance in reducing the relays total operation damage, avert the unintentional feeder disconnections or
time. Although a great research attention is dedicated to the nuisance DG tripping, and further improving of the reliability
conventional DOCRs, there are quite a few researches metrics.
conducted on dual-setting DOCRs. As the first attempt in [25], In brief, the main contributions of this paper are:
application of dual-setting DOCRs was proposed to protect the  A techno-economic optimal replacement level of dual-
radial distribution networks. It should be mentioned that despite setting DOCRs is attained;
deploying DOCRs to cope with bi-directional power flow issue  An efficient multi-objective protection coordination
in DG-mixed networks, the operation time is still long. By strategy is devised for DOCRs;
increasing DG penetration, the complexity of coordination  Non-standard coordination process is employed to provide
problem intensifies which results in higher operation times for higher flexibility;
relays. Accordingly, it is contemplated to deploy dual-setting  Remarkable reduction in total operation time of relays is
relays that can lower the protection operation time and prevent achieved.
these difficulties. More recently, application of dual-setting The established model lies within a non-linear programming
DOCRs has been interrogated in protecting multiple source fashion which is tackled based on PSO algorithm.
meshed distribution networks [26]. In [27], dual-setting DOCRs
were deployed for protecting a micro-grid capable of operating II. BASIC PRINCIPLE OF DUAL-SETTING DOCRS
in both grid-connected and islanded modes. In all of the
preceding studies, a sensible reduction has been reported in total A. Inverse-Time Characteristics of Dual-Setting DOCRs
operation time of relays. However, the conventional relays are A dual-setting DOCR provides individual and independent
completely replaced with dual-setting DOCRs with no attention protection for both fw/p and rv/b directions. General
to the economic burden of the new scheme. Not surprisingly, representation of these characteristics is:
deployment of dual-setting DOCRs reduces the operation time A
t  TDS  ( ) (1)
more significantly at the first stages of the replacement levels. IF B
( ) 1
However, the reduction in operation time saturates as the Ip
penetration of dual-setting DOCRs grows. Such an important
where, t represents operation time of each relay, IF denotes
point is overlooked in the previous studies since all the relays
fault current, and A and B are the coefficients of time-inverse
are supposed to be dual-setting DOCRs. This is while, an characteristics. In these relays, independency of the forward
optimal number of these relays in conjunction with the and reverse characteristics relaxes some of the coordination
conventional ones could offer a fast protection task in a cost-
constraints. This feature is the first contribution of dual-setting
effective manner [28]. Devising efficient models for optimal DOCRs in granting efficient and fast response strategies. Also,
deployment of dual-setting DOCRs is of curial significance in standard coordination strategies, the coefficients A and B in
which has not yet been covered in the existing literature.
(1) are assumed to be constant parameters; however in non-
This paper determines an optimal replacement level of
standard coordination approach, A and B are considered as
conventional DOCRs with dual-setting ones. A multi- variables to be optimized besides TDS and Ip. This
objective optimization model is proposed which aims at
countermeasure assists in achieving further reduction in relays
minimizing total operation time of relays through deployment
total operation time.
of optimal number of dual-setting DOCRs. Here, the number
of deployed relays implicitly represents the cost including B. Dual-Setting DOCRs versus Conventional DOCRs
purchase, installation, communication links, and other The protection coordination process mainly refers to
components. Since some of conventional DOCRs are replaced determination of primary and backup pair relays and provision
with dual-setting DOCRs, an efficient protection coordination of enough time difference between their operation times in
strategy is proposed to handle coordination problem in case of different fault conditions. Deploying diverse combinations of
attending both conventional and dual-setting DOCRs. To take dual-setting and conventional DOCRs end in different pair of
the most advantage of dual-setting DOCRs in reducing the primary and backup relays. For the sake of clarity, consider
overall operation time, non-standard coordination process is two scenarios illustrated in Fig. 1. In Fig. 1(a), where all relays
employed. Here, we add a new level of flexibility to the are considered as the conventional DOCRs, for a typical fault
coordination process by optimizing the time-current at point A, R1 and R6 are the primary set and R3 and R4
characteristic of the relays. To tackle the proposed multi- operate as the backup units. The second scenario, shown in
objective optimization model, augmented ε-constraint Fig. 1(a), deals with mixed deployment of conventional and
approach is deployed and Pareto-optimal solution regarding dual-setting relays. In this case, R2 and R6 are dual-setting
the total operation time of relays and the number of deployed relays and the remaining relays are conventional DOCRs. For
dual-setting DOCRs is attained. Next, a fuzzy decision making a similar fault at point A, R1 and R6 are the primary relays. As
process is utilized to determine the compromised solution in R2 is a dual-setting DOCR, the reverse direction of R2 is the
terms of these conflicting objectives. The main feature of the backup relay for R1 and backup of R6 is R4.
proposed model is achieving remarkable reduction in the total What should be emphasized is that the communication link

1551-3203 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2018.2821175, IEEE
Transactions on Industrial Informatics
3

among the relays is deployed to minimize the interrupted both C and D are subsets of I and
section by fulfilling selectivity constraints. For instance, CD  I (5)
consider Fig. 1(b) where R1 is the primary relay and the CD  0 (6)
reverse direction of R2 is the backup. As can be seen, R3 In regard of dual-setting DOCRs, individual characteristics are
might also be triggered in the backup path for the fault at point employed for both forward and reverse directions. That is:
A. This is while; R3 is not coordinated with R1 which might
Adfw
lead to over-interruption. To avert so, a blocking signal is sent t dfw, f TDS dfw ( ); d D (8)
IFd ,f Bdfw
by R2 to R3 through a communication link; accordingly, R3 is ( fw )  1
restrained until operation of R2. Ipd
R6
R1 Line 1 R6 R1
Line 1 Adrv
t drv, f TDS drv ( ); d D (9)
A fault A fault IFd ,f B drv
( ) 1
Ipdrv
R2 R5 R2 R5
R3 R4 R3 R4 Security of the protection coordination obligates existence
:DOCR :Dual-setting DOCR
of enough time difference, referred to as the coordination time
(a) (b) interval (CTI) between operation of the main relay and its
Fig. 1. Different deployment scenarios: a) all relays are conventional backup. This requirement is formulated by (10)-(13) for
DOCRs, b) conventional and dual-setting DOCRs are mixed. different combinations of conventional and dual-setting
It is worth mentioning that deployment of DOCRs is a DOCRs:
common practice to cope with bi-directional power flow issue t kfw ,rv  t bdrv , k  t dfw, k  CTI  0; d  D , bd  BD (10)
in DG-integrated distribution systems. The main concern with
conventional DOCRs is that their operation time is long which t kfw ,b  t bcb , k  t dfw, k  CTI  0; d  D , bc  BC (11)
could bring about numerous difficulties. In the proposed t kp ,rv  t bdrv ,k  t cp,k  CTI  0; c C ,bd  BD (12)
method, some of the conventional DOCRs are replaced with
dual-setting ones for the sake of reducing total operation time t k
p ,b
t b
bc , k t p
c ,k  CTI  0; c C ,bc  BC (13)
of the relays. Accordingly, bi-directional power flow issue is
resolved as all of the deployed relays are directional. For k th pairs relay, t k denotes the operation time difference
with reference to CTI. In these constraints, bc and bd count
III. THE PROPOSED DEPLOYMENT AND COORDINATION the number of backup conventional DOCRs and backup dual-
METHODOLOGY setting DOCRs, respectively. As well, BC and BD indicate the
A. Objective Functions and Constraints sets of backup conventional DOCRs and backup dual-setting
DOCRs, respectively. Both BC and BD are the subsets of
The main goal is to compromise two objective functions; overall backup relays set, denoted by BR. Mathematical
the first one deals with reducing the total operation time of representations of these constraints are as follows.
relays and the second one decreases the number of deployed
dual-setting DOCRs. As a general wisdom in protection
BC  BD  BR (14)
engineering, significance of a fast backup protection is equal BC  BD  0 (15)
to (or even more than) the primary one [26], [27]. Hence, BR  I (16)
concurrent minimization of the primary and backup relays
operation times is aimed here: As each dual-setting DOCR protects both forward and reverse
directions, number of optimization variables is twice the
  
F operating time     t ifw, f / p   t irv, f /b, s   (2) conventional types, i.e., each of the optimization variables are
f F 
 i I  s S i   assessed in both forward and reverse directions as follows.
where and denote the operation times of fw/p and rv/b units TDS min  TDS dfw  TDS max ; d  D (17)
based on dual-setting/conventional DOCRs, respectively.
Moreover, f, i, and s are the indices for fault point, relay, and TDS min  TDS rv
d  TDS max ; d  D (18)
backup relay, respectively. Also, F, I, and S represent the sets of TDS min  TDS c  TDS max ; c C (19)
fault points, all relays, and backup relays.
The second objective aims at decreasing the number of dual- Ipmin  Ipdfw  Ipmax ; d  D (20)
setting DOCRs. That is: Ip min  Ip rv
 Ipmax ; d  D (21)
d

F Dual  Deployment   u i (3) Ipmin  Ipc  Ipmax ; c C (22)


i I

0 if relay(i)is conventional Constraints (5)-(22) are the main basis for extracting the
ui   (4) required settings of standard coordination strategies. In non-
1 if relay(i)is dual-setting standard coordination strategies, coefficients A and B in (7)-
The relays which are conventional types belong to set C and (9) are also specified in the optimization process. Associated
dual-setting relays are defined with set D. Needless to remind that constraints are elaborated by (23)-(28).

1551-3203 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2018.2821175, IEEE
Transactions on Industrial Informatics
4

Amin  Adfw  Amax ; d  D (23) process is used to adopt the final solution [31], [32]. For each
objective function, a linear membership function is assumed.
Amin  Adrv  Amax ; d  D (24) 1 ; Fj , Np  Fj ,min

Amin  Ac  Amax ; c C (25)   Fj ,max  Fj , Np  (33)
 j , Np    ; Fj ,min  Fj , Np  Fj ,max
B min  B fw
 B max ; d  D (26)   Fj ,max  Fj ,min 
d
0 ; Fj . Np  Fj ,max

B min  B drv  B max ; d  D (27) th
where Fj , Np denotes the value of j objective function pertaining
B min  Bc  B max ; c C (28) to Npth Pareto-optimal solution. Moreover,  j , Np represents the
In a case where some relays coefficients are adjustable and corresponding degree of membership function. Considering total
the other ones are fixed, the adjustable ones are determined q+1 solutions, a normalization process is performed on the
optimally in the optimization process and the remaining ones obtained membership function values.
are predetermined fixed values.
  j  j , Np
n

The operation time of each DOCR should satisfy the j 1


Np  (34)
   j  j , Np
m n
minimum and maximum permissible limits represented by
Np 1 j 1
(29)-(30). These constraints guarantee a fast and reliable
where the number of Pareto-optimal solutions and objective
protection scheme.
functions are represented by m and n, respectively. Also, weight
t min  t dfw  t max ; d  D (29) of jth objective function is denoted by ωj. Eventually, the
t min  t c  t max ; c C (30) compromised Pareto-optimal result is recognized by the
solution with maximum membership function of µNp.
In the proposed method, overall operation time of relays is To maintain diversity of the swarm and overcome premature
aimed to be minimized. The operation time of some relays convergence, the social component and the cognitive
might increase though decreasing overall operation time. To component in the updating velocity equation are supposed to be
avert such solutions and guarantee operation time reduction in dependent [33]. Positive linear correlation has better
all relays, operation time of the primary relays are restricted to performance than completely negative linear correlation and
an upper limit which is a common practice in the literature [20]. independence [34]. Therefore, in the conducted study, based on
B. Optimization Approach the presented results in [34], positive linear correlation is
considered.
The PSO algorithm is utilized as the optimization engine.
Cutting down the total operation time of relays necessitates
IV. NUMERICAL STUDIES: RESULTS AND DISCUSSIONS
further deployment of dual-setting DOCRs. This observation
means that the objective functions act conversely. To reach a A. Test System
trade-off between these two conflicting objectives, the augmented Extensive numerical studies are carried out to evaluate
ε-constraint method [29], [30] is utilized here. In this method, performance of the proposed approach. A typical 9-bus multiple
F operating time is considered as the main objective; whereas, for the source meshed distribution system is simulated in DIgSILENT
second objective, its variation range is merely calculated. Next, PowerFactory 14.3.1 and analyzed in depth. Fig. 2 depicts
the recorded range is divided into q equal intervals. Accordingly, single-line diagram of the test system whose main data is
there are q+1 grid points attained for F Dual Deployment . These q+1 available in [35]. This system contains 12 DOCRs which are
sub-problems are evaluated as follows. conventional relays, basically. However, some of them would be
 S  replaced by dual-setting DOCRs. Note that CTI takes a value
Min  F operating time    ( 2 )  (31) between 0.2 and 0.5 sec, the minimum values of coefficients A
 r2 
and B are assumed to be 0.14 and 0.02, and their maximum
Subject to: limits are assigned as 13.5 and 1, respectively [25]. Also, the
 F Dual  Deployment  S2   2 , S2  R  minimum and maximum limits of relays operation time are
 Dual  Deployment Dual  Deployment
  F Dual  Deployment  ( Fmax  Fmin
)  Np Line 3 Line 4

2 max
q R5 F3 R6 R7 F4 R8
 Np  0, 1, ... , q G G
 R4 R9
2 7 F2 8 Line 5 F5 9 3
 All other constraints Line 2
(32) R3 R10
5 6
where,  is a small value, S2 is an interval variable, r2 is the R2 R11
Line 1 Line 6
variation range of the number of dual-setting DOCRs obtained F1 F6
R1 R12
as FmaxDual  Deployment Dual  Deployment
 Fmin ,  2 is the Npth range of 4
F Dual Deployment , and q represents the number of equal parts
1
achieved for the variation range of F Dual Deployment .
G

Fig. 2. Single-line diagram of the investigated 9-bus test system.


Once the Pareto-optimal solution is obtained, a fuzzy

1551-3203 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2018.2821175, IEEE
Transactions on Industrial Informatics
5

considered as 0.1 and 2.5 sec, respectively [18]. Ultimately, it is method is executed. The possible number of dual-setting
assumed that DGs provide fault-ride-through requirement of DOCRs are 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, and 12; so, their
the deployed current-based relays. A cornerstone requirement corresponding percentage would be equal to 0, 8.33, 16.67, 25,
of current-based relays is that their accurate operation 33.33, 41.67, 50, 58.33, 66.67, 75, 83.33, 91.67, and 100%,
necessitates fault current provision by resources for a given respectively. In this way, 13 sub-problems should be evaluated.
period of time. This is referred to as the fault ride-through The obtained Pareto-optimal front is shown in Fig. 3. Both
requirement. Here, it is assumed that current-based protection objectives are treated with a similar importance. Table II
is realized by resources through enough ingredients. represents the obtained results. In Table II, the maximum
Availability of this requirement is a common assumption for membership degree denotes the best compromise solution. As
protection device deployment and setting [36]. can be seen, the optimal deployment percentage of dual-setting
DOCRs is obtained as 41.67%. The solution vector renders R2,
B. Results
R7, R8, R10, and R11 as dual-setting DOCRs. Referring to
The optimal deployment of dual-setting DOCRs is Table II, the proposed procedure decreases the operation time of
assessed in two different scenarios. The considered scenarios relays from 17.8428 to 13.2021 sec. At higher replacement
investigate deployment of dual-setting DOCRs instead of both levels, there is a slight decrease in total operation time of relays.
kind of relays which can or cannot support non-standard This notice is valid for both the fw/p and the rv/b DOCRs. Thus,
coordination process. In the first scenario, the problem is a saturated behavior is observed in protection time reduction
handled based on the standard characteristics while in the (see Fig. 3). For the best compromised solution with 41.67%
second one, the non-standard characteristics are explored. The replacement level, the optimal settings are reported in Table III.
coordination process could be tackled considering single-point
or multi-point faults. In the former way, a three-phase solid 18
fault is considered at the midpoint of each protected line. In

Sum of operation
time of relays (s)
the latter one, in addition to the midpoints, the faults at near- 16
end and far-end of the protected line are also investigated.
1. Scenario I: Standard Coordination Process based on 14
Single-Point Approach
In the base case, all of the DOCRs are conventional relays. 12
0 8.3 16.7 25 33.3 41.7 50 58.3 66.7 75 83.3 91.7 100
Here, TDS and Ip denote the optimization variables and standard % of penetration of dual setting relays (%)
values are assigned to A and B. The obtained settings are Fig. 3. Pareto-optimal front in Scenario I.
reported in Table I. The total operation time of relays in primary TABLE III
and backup processes is 17.8428 sec. To determine the optimal OPTIMAL SETTINGS OF RELAYS IN SCENARIO I WITH 41.67% DEPLOYMENT LEVEL
deployment of dual-setting DOCRs, the augmented ε-constraint Relay
Parameters
No. TDS fw
Ip fw TDS rv Ip rv
TABLE I
OPTIMAL SETTINGS OF CONVENTIONAL DOCRS IN SCENARIO I 1 0.1000 0.1265 - -
Relay Parameters Relay Parameters 2 0.1000 0.1551 0.1913 0.1596
No. TDS Ip (kA) No. TDS Ip (kA)
3 0.1000 0.2178 - -
1 0.2168 0.14 7 0.212 0.0954
4 0.1217 0.2447 - -
2 0.1126 0.1811 8 0.329 0.065
3 0.1 0.2276 9 0.1862 0.1694 5 0.1344 0.2208 - -
4 0.1312 0.2617 10 0.111 0.1987 6 0.1087 0.2173 - -
5 0.2111 0.2123 11 0.1479 0.1203 7 0.1000 0.0986 0.2702 0.0964
6 0.1046 0.2438 12 0.2776 0.0856 8 0.1000 0.0649 0.1812 0.0653
9 0.1000 0.1694 - -
TABLE II 10 0.1000 0.1661 0.1430 0.1661
PARETO POINTS OF DEPLOYMENT OF DUAL-SETTING DOCRS IN SCENARIO I
11 0.1079 0.1007 0.2284 0.0981
Pareto Total Replacement
points operation time of dual-setting 1,Np 2,Np Np 12 0.1000 0.0825 - -
(Np) of relays (sec) DOCRs (%)
1 12.70 100.0 1.000 0.000 0.083 The operation time of the relays is given in Table IV. Not
2 12.70 91.67 0.998 0.083 0.086 only the coordination constraints are met, the operation times of
3 12.72 83.33 0.996 0.166 0.088 individual relays are decreased compared to those reported in
4 12.72 75.00 0.995 0.250 0.090
5 12.80 66.67 0.980 0.333 0.091
Table II. Moreover, Table V reports total operation time of the
6 12.97 58.33 0.947 0.416 0.091 relays for different values of CTI. As can be seen, the optimal
7 13.09 50.00 0.923 0.500 0.091 solution remains almost unchanged for different CTIs. The
8 13.20 41.67 0.902 0.583 0.091
reason is that time characteristics of the overcurrent relays are
9 13.50 33.33 0.843 0.666 0.089
10 15.34 25.00 0.486 0.750 0.061 parallel inverse curves representing different time dial settings.
11 15.77 16.67 0.403 0.833 0.057 Therefore, difference between the curves remains almost fixed
12 16.33 8.330 0.293 0.916 0.050 for different values of CTI. If the coordination constraints are
13 17.84 0.000 0.000 1.000 0.027
fulfilled for a specific CTI, it can be also assured for other
Best
13.20 41.67% 0.903 0.583 0.092 values.
solution

1551-3203 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2018.2821175, IEEE
Transactions on Industrial Informatics
6

TABLE IV TABLE VII


OPERATION TIME IN SCENARIO I WITH 41.67% DEPLOYMENT LEVEL PARETO POINTS OF DEPLOYMENT OF DUAL-SETTING DOCRS IN SCENARIO II
Operation times of relays in sec. Total operation Replacement of
Pareto points
Fault (fw/p= forward/primary, rv/b = reverse/backup) time of relays dual-setting 1,Np 2,Np Np
(Np)
location (sec) DOCRs (%)
fw/p rv/b
1 43.1883 100 1 0 0.0807
R1: 0.2613 R11: 0.5613 2 43.2236 91.67 0.9993 0.0833 0.0829
F1
R2: 0.4264 R4: 0.7264 3 43.3034 83.33 0.9978 0.1667 0.085
R3: 0.5926 R2: 0.8926 4 43.323 75 0.9974 0.25 0.0872
F2 5 43.8963 66.67 0.9864 0.3333 0.0886
R4: 0.3870 R6: 0.6870
6 45.5341 58.33 0.9549 0.4167 0.0883
R5: 0.4462 R3: 0.9293
F3 7 45.8644 50 0.9486 0.5 0.09
R6: 0.5279 R7: 0.8279
8 49.1093 41.67 0.8863 0.5833 0.0872
R7: 0.2866 R5: 0.5866 9 50.9077 33.33 0.8517 0.6667 0.0867
F4
R8: 0.2221 R10: 0.7476 10 60.4712 25 0.668 0.75 0.0741
R9: 0.3447 R8: 0.6447 11 61.651 16.67 0.6454 0.8333 0.0745
F5
R10: 0.4129 R11: 0.7129 12 80.1944 8.33 0.2892 0.9167 0.048
R11: 0.3792 R10: 0.6792 13 95.2521 0 0 1 0.0269
F6 Best solution 45.8644 50% 0.9486 0.5 0.09
R12: 0.2240 R2: 0.6957
Total 4.5109 8.6912 TABLE VIII
operation     OPTIMAL SETTINGS OF RELAYS IN SCENARIO II WITH 50% DEPLOYMENT
time (sec) LEVEL
0.3759 0.1159 0.7243 0.1120
Relay Parameters
TABLE V
SIMULATION RESULTS FOR DIFFERENT CTI VALUES
No. TDS fw Ip fw TDS rv Ip rv
1 0.1 0.1265 0.3221 0.1454
Replacement
F operation time NP
Np of dual-setting 2 0.1005 0.162 - -
DOCRs (%) 0.2 0.3 0.4 0.2 0.3 0.4 3 0.1 0.2178 - -
4 0.1 0.2343 0.7908 0.3064
1 100.0 12.22 12.70 13.92 0.058 0.083 0.061
5 0.1 0.2057 0.2048 0.2119
2 91.67 12.22 12.70 13.92 0.063 0.086 0.067 6 0.1119 0.2087 - -
3 83.33 12.22 12.72 13.92 0.068 0.088 0.072 7 0.1 0.0935 - -
4 75.00 12.22 12.72 13.92 0.073 0.090 0.077 8 0.1 0.0649 - -
5 66.67 12.22 12.80 14.04 0.078 0.091 0.081 9 0.1701 0.1748 0.3434 0.2155
6 58.33 12.24 12.97 14.24 0.082 0.091 0.084 10 0.1 0.1661 0.4753 0.1918
7 50.00 12.27 13.09 14.45 0.086 0.091 0.087 11 0.1414 0.0946 - -
8 41.67 12.37 13.20 14.88 0.088 0.092 0.089 12 0.1 0.0825 0.975 0.0964
9 33.33 12.60 13.50 15.90 0.085 0.089 0.085 TABLE IX
10 25.00 12.70 15.34 16.50 0.086 0.061 0.084 OPERATION TIME IN SCENARIO II WITH 50% DEPLOYMENT LEVEL
11 16.67 12.99 15.77 18.14 0.081 0.057 0.076 Fault point
12 8.330 13.00 16.33 18.59 0.086 0.050 0.077 Pair relays
Near-end Mid-point Far-end
13 0.000 13.97 17.84 20.93 0.058 0.027 0.061
PR1 BR2 t primary tbackup t primary tbackup t primary tbackup
Best
solutio 41.67% 12.379 13.20 13.918 0.088 0.092 0.089 1 12 0.2309 0.5310 0.2613 0.7021 0.2906 0.9693
n 2 4 0.3953 0.6954 0.4404 0.7983 0.4990 0.9422
3 1 0.3799 0.6800 0.5926 1.0205 0.6332 1.0827
2. Scenario II: Standard Coordination Process based on 4 5 0.2923 0.5923 0.3116 0.6818 0.5008 2.1521
Multi-Point Approach 5 4 0.2825 0.6332 0.3209 0.9293 0.3594 1.6777
6 8 0.4715 0.7715 0.5283 0.8329 0.5923 0.8969
Single-point coordination approach may depreciate accuracy 7 5 0.2524 0.5526 0.2802 0.6554 0.3102 0.7829
of protection coordination [10], [13], more specifically for the 8 9 0.1998 0.5181 0.2221 0.7476 0.2429 1.3335
faults that occur in far-end of a protected line. To enhance the 9 7 0.4636 0.7636 0.5960 1.0209 0.7411 1.4746
10 12 0.3333 0.6333 0.4129 0.7557 0.5181 0.9043
coordination accuracy, multi-point coordination strategies are 11 10 0.4410 0.7411 0.4813 0.8378 0.5310 0.9670
proposed such as two-point approach in [13] and three-point 12 1 0.1990 0.4990 0.2240 0.7310 0.2470 1.2777
approach in [26]. In this scenario, the three-point coordination TOT 3
3.9415 7.6111 4.6716 9.7133 5.4656 14.4609
method with near-end, mid-point, and far-end faults is adopted. µ σ µ σ µ σ µ σ µ σ µ σ
(sec)
0.33 0.1 0.63 0.097 0.39 0.138 0.81 0.124 0.45 0.16 1.2 0.4
The obtained results associated with 0% deployment of dual- 1 Primary Relay, 2 Backup Relay, 3 Total Operation Time

TABLE VI
OPTIMAL SETTINGS OF CONVENTIONAL DOCRS IN SCENARIO II setting DOCRs are reported in Table VI.
Parameters Parameters
As can be seen, the total operation time of relays is 95.2521
Relay Relay
sec. Having evaluated the multi-objective approach on all
No. TDS Ip (kA) No. TDS Ip (kA) discretized intervals, Table VII specifies Pareto points and the
1 0.3447 0.1363 7 0.3443 0.0935 best compromised solution which has 50.0% deployments level
2 0.2544 0.1551 8 0.4868 0.0649 of dual-setting DOCRs. The solution vector includes R1, R4,
3 0.1925 0.2178 9 0.2819 0.1748 R5, R9, R10, and R12. Doing so, the operation time of relays is
4 0.2538 0.2357 10 0.2458 0.1661 decreased from 95.25 to 45.86 sec. In comparison to the
5 0.3253 0.2057 11 0.2976 0.0946
obtained results in single-point approach, the three-point
6 0.1978 0.2079 12 0.4624 0.0840
method results in an increased percentage of dual-setting

1551-3203 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2018.2821175, IEEE
Transactions on Industrial Informatics
7

TABLE XI
DOCRs from 41.66% to 50% for assuring the anticipated PARETO POINTS OF DEPLOYMENT OF DUAL-SETTING DOCRS IN SCENARIO
performance of the protection system. At 50% deployment III
level of dual-setting DOCRs, the optimal settings are Pareto points
Total operation Replacement of
time of relays dual-setting 1,Np 2,Np Np
reported in Table VIII. The operation time of relays at (Np)
(sec) DOCRs (%)
different fault points are also given in Table IX. At this 1 6.0000 100 1.0000 0 0.0645
penetration level, the total operation time of the relays is 2 6.0308 91.67 0.9933 0.0833 0.0695
attained as 45.8644 sec. 3 6.1407 83.33 0.9695 0.1667 0.0733
3. Scenario III: Non-Standard Coordination Process based 4 6.2314 75 0.9499 0.2500 0.0774
on Single-Point Approach 5 6.5387 66.67 0.8833 0.3333 0.0785
Here, in addition to the typical variables TDS and Ip, 6 6.8666 58.33 0.8122 0.4167 0.0793
coefficients of A and B are also included in the optimization 7 7.0686 50 0.7685 0.5000 0.0819
process. Accordingly, the conventional DOCRs are modeled 8 7.2412 41.67 0.7311 0.5833 0.0848
with four optimization variables whereas the dual-setting 9 7.5080 33.33 0.6733 0.6667 0.0865
DOCRs are expressed with eight variables. At first, the 10 8.0632 25 0.5530 0.7500 0.0841
deployment level of dual-setting DOCRs is assumed to be 11 8.7361 16.67 0.4072 0.8333 0.0801
12 9.4453 8.33 0.2535 0.9167 0.0755
0% for which the optimal settings of relays are given in
13 10.6152 0 0 1.0000 0.0645
Table X. In this case, the total operation time of relays is
Best solution 7.5080 33.33 0.6733 0.6667 0.0865
attained as 10.6153 sec. As obvious, a significant reduction
is attained compared to that of the standard approach, say TABLE XII
17.8428 sec at first scenario (see Table II). Therefore, non- OPTIMAL SETTINGS OF RELAYS IN SCENARIO III WITH 33.33% DEPLOYMENT
LEVEL
standard strategy provides more flexibility in procuring
Parameters
efficiently coordinated and fast protection strategy. Table XI Relay fw
represents the obtained results for each of sub-problems. fw Ip
No. TDS TDS rv Ip rv (kA) A fw A rv B fw B rv
(kA)
Here, the optimal deployment level of dual-setting DOCRs is
1 0.3696 - 0.1498 - 1.8954 - 0.8522 -
attained as 33.33% which means 4 replacements of 11.236
2 0.1004 0.1000 0.1885 0.1730 4.3488 0.9215 0.8641
conventional DOCRs with dual-setting ones. For the best 4
compromised solution, R2, R6, R7, and R11 are determined 3 0.1950 - 0.2661 - 1.4945 - 1.0000 -
4 0.1409 - 0.2502 - 5.6486 - 0.8812 -
as dual-setting DOCRs. The obtained solution results in 5 0.2660 - 0.2312 - 2.3581 - 0.9830 -
reduction of relays operation time from 17.8428 to 7.5080 6 0.4055 0.1769 0.2751 0.2487 1.2548 9.6330 0.8887 0.9873
sec. Comparison of the results with those of Scenario I 7 0.2391 0.4926 0.1200 0.0963 3.6642 4.2575 0.7201 0.6887
8 0.3398 - 0.0802 - 4.4693 - 0.9784 -
reveals that although Scenario III has fewer deployment of 9 0.4272 - 0.1925 - 2.2418 - 0.8443 -
dual-setting DOCRs, it leads to 43.13% extra reduction in 10 0.1402 - 0.1772 - 3.9005 - 0.9987 -
total operation time. This observation is supported by higher 11 0.1493 0.2135 0.1200 0.1066 5.7826 5.4004 0.9980 0.6058
12 0.4945 - 0.0919 - 3.0251 - 0.9475 -
flexibility in adjusting the inverse-time characteristics by
relaxing coefficients of A and B to be optimized. At 33.33% TABLE XIII
deployment of dual-setting DOCRs, optimal settings of OPERATION TIME IN SCENARIO III WITH 33.33% DEPLOYMENT LEVEL
optimization variables are reported in Table XII for both Operation times of relays in sec.
Fault (fw/p= forward/primary, rv/b = reverse/backup)
forward and reverse directions. Besides, the operation time location
fw/p rv/b
of relays are given in Table XIII for both primary and R1: 0.1 R11: 0.4
F1
backup protections. R2: 0.1617 R4: 0.4617
TABLE X R3: 0.1787 R2: 0.4787
F2
OPTIMAL SETTINGS OF CONVENTIONAL DOCRS IN SCENARIO III R4: 0.1435 R6: 0.4435
Parameters R5: 0.1 R3: 0.4
Relay F3
R6: 0.2725 R7: 0.5731
No. TDS Ip (kA) A B R7: 0.2285 R6: 0.5285
F4
1 0.1281 0.158 7.5765 0.6262 R8: 0.1 R10: 0.4
2 0.2508 0.1914 1.9936 0.8032 R9: 0.2507 R7: 0.5507
F5
3 0.2375 0.2453 2.4743 0.9748 R10: 0.1383 R11: 0.4448
R11: 0.1764 R9: 0.4764
4 0.1721 0.2599 6.0726 0.9804 F6
R12: 0.1 R2: 0.4
5 0.1623 0.22 3.2781 0.4182
Total 1.9503 5.5574
6 0.1012 0.2431 7.338 0.9856
operation    
7 0.2442 0.1001 5.4102 0.7519 time (sec) 0.1625 0.0612 0.4631 0.0611
8 0.2031 0.0782 7.236 0.5103
9 0.2057 0.1883 3.3061 0.5930
In this scenario, three-point coordination is adopted. The
10 0.1747 0.1895 5.0958 0.9906
obtained results associated with 0% deployment of dual-
11 0.1917 0.1074 7.0895 0.9549
setting DOCRs are reported in Table XIV. As can be seen,
12 0.5783 0.0915 8.0303 0.9927
total operation time of the relays is 67.40 sec. Having
4. Scenario IV: Non-Standard Coordination Process based
evaluated the multi-objective approach on all discretized
on Multi-Point Approach
intervals, Table XV specifies Pareto points and the best

1551-3203 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2018.2821175, IEEE
Transactions on Industrial Informatics
8

TABLE XIV compromised solution with 41.67% deployment level of


OPTIMAL SETTINGS OF CONVENTIONAL DOCRS IN SCENARIO IV dual-setting DOCRs. The solution vector includes R1, R2,
Relay Parameters R6, R7, and R12 as the optimal replacement candidates of
No. TDS Ip (kA) A B
1 1.1000 0.1725 3.1330 1
dual-setting DOCRs to reach the minimum operation times.
2 1.1000 0.2115 1.0728 1 The operation time of relays is decreased from 67.4 to 35.8
3 0.6443 0.2970 1.1531 1 seconds. Comparing the obtained results with those of the
4 1.1000 0.3195 1.1245 1
5 1.0573 0.2805 2.2244 1 single-point approach, it can be deduced that three-point method
6 1.0209 0.2735 0.8662 1 results in an increased percentage of dual-setting DOCRs
7 1.1000 0.1275 2.6520 1 deployment from 33.33% to 41.67% to assure the anticipated
8 1.1000 0.0885 6.0869 1
9 0.7872 0.2310 2.2551 1 performance of protection system. At 41.67% deployment of
10 0.9083 0.2265 1.2230 1 dual-setting DOCRs, optimal settings for both forward and
11 1.0245 0.1290 1.9773 1
12 0.9113 0.1125 6.1895 1 reverse directions are reported in Table XVI. The operation time
of the relays at different fault points are also given in Table
TABLE XV
PARETO POINTS OF DEPLOYMENT OF DUAL-SETTING DOCRS IN SCENARIO IV XVII. As can be seen, the obtained results on different scenarios
Replacement of confirm outperformance of the proposed approach in cutting
Pareto points Total operation
dual-setting 1,Np 2,Np Np down the total operation time of relays and attaining the
(Np) time of relays (sec)
DOCRs (%)
1 25.8572 100 1.0000 0 0.0650
minimum number of dual-setting DOCRs deployments.
2 26.0200 91.67 0.9961 0.0833 0.0702
3 26.9929 83.33 0.9727 0.1667 0.0741 C. Results Validations
4 27.5322 75 0.9597 0.2500 0.0786 IEEE 30-bus test system with 62 pair relays is used to test
5 29.1406 66.67 0.9210 0.3333 0.0815
6 30.7925 58.33 0.8812 0.4167 0.0844 scalability of the proposed approach. In this manner, standard
7 32.8956 50 0.8306 0.5000 0.0865 coordination process with single-point approach based on the
8 35.8481 41.67 0.7595 0.5833 0.0873
9 39.8177 33.33 0.6640 0.6667 0.0865
conducted power flow and short circuit analyses in [37] is
10 44.1700 25 0.5592 0.7500 0.0851 performed. Saturated behavior is observed in protection time
11 58.0008 16.67 0.2264 0.8333 0.0689
12 62.6090 8.33 0.1155 0.9167 0.0671 TABLE XVIII
13 67.4058 0 0 1.0000 0.0650 PARETO POINTS OF DEPLOYMENT OF DUAL-SETTING DOCRS IN IEEE 30 BUS
Best solution 39.8177 41.67 0.7595 0.5833 0.0873 SYSTEM
Total operation Replacement of
Pareto points
TABLE XVI time of relays dual-setting 1, Np 2, Np Np
(Np)
OPTIMAL SETTINGS OF RELAYS IN SCENARIO IV WITH 41.67% DEPLOYMENT (sec) DOCRs (%)
LEVEL 1 52.236 100 1 0 0.02
Parameters 2 52.236 97.37 1 0.0263 0.0206
Relay 3 52.236 94.74 1 0.0526 0.0211
No. TDS fw TDS rv Ip fw Ip rv (kA) A fw 4 52.337 92.11 0.9989 0.0789 0.0216
A rv B fw B rv
(kA) 5 52.381 89.47 0.9984 0.1053 0.0221
1 0.1000 0.100 0.1265 0.1265 0.140 0.1400 0.0445 0.0210 6 52.401 86.84 0.9981 0.1316 0.0226
2 0.1000 0.164 0.1551 0.1551 0.140 0.1638 0.0732 0.0200 7 52.539 84.21 0.9966 0.1579 0.0231
3 0.1000 0.100 0.2178 0.2178 0.140 0.1400 0.0244 0.0237 8 52.541 81.58 0.9966 0.1842 0.0237
4 0.8751 0.100 0.2648 0.2343 1.148 0.1400 1.0000 0.0261 9 52.989 78.95 0.9915 0.2105 0.0241
5 1.0516 0.100 0.2805 0.2057 1.505 0.1400 1.0000 0.0247 10 55.224 76.32 0.9663 0.2368 0.0241
6 0.1000 0.111 0.2079 0.2079 0.140 0.1400 0.0800 0.0200 11 55.247 73.68 0.9661 0.2632 0.0246
7 0.1000 0.143 0.0935 0.0935 0.140 0.1400 0.0485 0.0200 12 56 71.05 0.9576 0.2895 0.025
8 1.1000 0.185 0.0885 0.0649 4.033 0.1400 1.0000 0.0200 13 56.1 68.42 0.9565 0.3158 0.0255
9 1.1000 0.100 0.1913 0.1694 1.769 0.1400 1.0000 0.0204 14 56.1 65.79 0.9565 0.3421 0.026
10 0.1000 0.173 0.1661 0.1661 0.140 0.2720 0.0201 0.0491 15 56.1 63.16 0.9565 0.3684 0.0266
11 1.1000 0.243 0.0946 0.0946 1.767 0.1400 0.9659 0.0200 16 56.119 60.53 0.9562 0.3947 0.0271
12 0.1000 0.115 0.0825 0.0825 0.140 0.1400 0.0385 0.0200 17 56.292 57.89 0.9543 0.4211 0.0276
18 56.795 55.26 0.9486 0.4474 0.028
TABLE XVII 19 57.971 52.63 0.9354 0.4737 0.0282
OPERATION TIME IN SCENARIO IV WITH 41.67% DEPLOYMENT LEVEL 20 59.683 50 0.9161 0.5 0.0284
Fault point 21 65.267 47.37 0.8532 0.5263 0.0276
Pair relays 22 65.701 44.74 0.8483 0.5526 0.0281
Near-end Mid-point Far-end
23 67.258 42.11 0.8307 0.5789 0.0283
PR1 BR2 t primary t backup t primary t backup t primary t backup 24 68.925 39.47 0.8119 0.6053 0.0284
1 12 0.1000 0.4000 0.1136 0.6866 0.1268 1.1743 25 69.652 36.84 0.8038 0.6316 0.0288
2 4 0.1000 0.4000 0.1116 0.5165 0.1267 0.6907 26 77.373 34.21 0.7168 0.6579 0.0275
3 1 0.3107 0.6107 0.4853 0.8764 0.5187 0.9225 27 78.15 31.58 0.708 0.6842 0.0279
4 5 0.1229 0.4229 0.1442 0.4862 0.2815 1.3764 28 78.802 28.95 0.7006 0.7105 0.0283
5 4 0.2187 0.5187 0.3041 0.7617 0.4000 1.3762 29 79.3 26.32 0.695 0.7368 0.0287
6 8 0.1000 0.4000 0.1126 0.4838 0.1268 0.5761 30 82.235 23.68 0.662 0.7632 0.0286
7 5 0.1000 0.4000 0.1115 0.5621 0.1238 0.7793 31 84.491 21.05 0.6365 0.7895 0.0286
8 9 0.2144 0.5144 0.3040 0.7422 0.4000 1.3239 32 91.276 18.42 0.5601 0.8158 0.0276
9 7 0.1910 0.4910 0.3548 0.6564 0.5649 0.9481 33 95.765 15.79 0.5095 0.8421 0.0271
10 12 0.3308 0.6308 0.4099 0.7441 0.5144 0.8782 34 101.159 13.16 0.4487 0.8684 0.0264
11 10 0.2649 0.5649 0.3236 0.7174 0.4000 0.9333 35 108.289 10.53 0.3684 0.8947 0.0253
12 1 0.1000 0.8637 0.1129 1.1217 0.1248 1.5471 36 114.493 7.89 0.2985 0.9211 0.0244
2.1533 6.2171 2.8881 8.3552 3.7084 12.5260 37 116.482 5.26 0.2761 0.9474 0.0245
TOT3 (sec) µ σ µ σ µ σ µ σ µ σ µ σ 38 121.31 2.63 0.2217 0.9737 0.024
0.18 0 0.51 0.19 0.24 0.02 0.69 0.03 0.3 0.03 1 0.1 39 140.981 0 0 1 0.02
1 Primary Relay, 2 Backup Relay, 3 Total Operation Time Best Solutions 69.65 36.84 0.8038 0.6316 0.0288

1551-3203 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2018.2821175, IEEE
Transactions on Industrial Informatics
9

reduction which is depicted in Fig. 4. Herein, Np is 38 and the [2] R. Minciardi and M. Robba, “A Bilevel approach for the stochastic
optimal operation of interconnected microgrids,” IEEE Trans. Autom.
Pareto optimum solutions are reported in Table. XVIII. As can Sci. Eng., vol. 14, no. 2, pp. 482-493, April 2017.
be seen, the best compromised solution represents 36.84% [3] Q. Kang, M. C. Zhou, J. An, Q. D. Wu “Swarm intelligence approaches to
deployment of dual-setting DOCRs. The solution vector optimal power flow problem with distributed generator failures in power
networks,” IEEE Trans. Autom. Sci. Eng, vol. 10, no. 2, pp. 343 353, 2013.
includes R4, R9, R12, R13, R14, R16, R18, R29, R30, R31,
[4] L. Huchel and H. H. Zeineldin. “Planning the coordination of directional
R32, R34, R37, and R38 as the optimal replacement overcurrent relays for distribution systems considering DG,” IEEE Trans.
candidates of dual-setting DOCRs to reach the minimum Smart Grid., vol. 7, no. 3, pp. 1642 1649, 2016.
operation times. [5] A. Khodaei, “Microgrid optimal scheduling with multi-period islanding
constraints.” IEEE Trans. Power Syst., vol. 29, no. 3, pp. 1383-1392, 2014.
[6] M. Singh, B. K. Panigrahi, and A. R. Abhyankar, “Optimal coordination of
directional over-current relays using teaching learning-based optimization
(TLBO) algorithm,” Int. J. Elec. Power., vol. 50, pp. 33 41, 2013.
[7] H. B. Elrafie, and M. R. Irving, “Linear programming for directional
overcurrent relay coordination in interconnected power systems with
constraint relaxation,” Electr. Pow. Syst. Res., vol. 27, no. 3, pp. 209 216,
1993.
[8] A. M. Othman, and A. Y. Abdelaziz, “Enhanced backtracking search
algorithm for optimal coordination of directional over-current relays
including distributed generation,” Electr. Pow. Compo. Sys., vol. 44, no.
Fig. 4. Pareto-optimal front in Scenario I. 3, pp. 278 290, 2016.
The conducted study is the initial attempt to deploy the dual- [9] C. A. C. Salazar, A. C. Enríquez, and S. E. Schaeffer. “Directional
overcurrent relay coordination considering non-standardized time
setting DCORs with techno-economic considerations. Here, bi- curves,” Electr. Pow. Syst. Res. vol. 122, pp. 42-49, 2015.
directional power flow issue in DG-integrated distribution [10] K. A. Saleh, H. H. Zeineldin, A. Al-Hinai, and E. F. El-Saadany, “Dual-
systems is resolved as all of the deployed relays are directional. setting characteristic for directional overcurrent relays considering
multiple fault locations,” IET Gener. Transm. Dis., vol. 9, no. 12, pp.
On the contrary, difficulties raised by system configuration 1332 1340, 2015.
alteration is still remained which is a common problem among [11] M. H. Costa, R. R. Saldanhaet, M. G. Ravetti, E. G. Carrano “Robust
all current-based protection approaches. To cope with such a coordination of directional overcurrent relays using a metaheuristic
algorithm.” IET Gener. Transm. Dis., vol. 11, no. 2, pp. 464-474, 2017.
problem, group setting-based protection coordination is [12] M. Y. Shih, A. C. Enríquez, T. Hsiao, L. M. T. Trevino, “Enhanced
proposed in the literature [38]. Taking the system configuration differential evolution algorithm for coordination of directional overcurrent
variations and accompanied alterations into account might be relays.” Electr. Pow. Syst. Res., vol. 143, pp. 365-375, 2017.
[13] M. N. Alam, B. Das, V. Pant, “A comparative study of metaheuristic
the subject of a possible future research study. optimization approaches for directional overcurrent relays coordination.”
Electr. Pow. Syst. Res., vol. 128, pp. 39-52, 2015.
V. CONCLUSION [14] V. N. Rajput, K. S. Pandya. “Coordination of directional overcurrent relays
in the interconnected power systems using effective tuning of harmony
The optimal deployment of dual-setting DOCRs in search algorithm.” Sustain Comput-infor, vol. 15, pp. 1-15, 2017.
multiple source meshed distribution networks was tackled in [15] A. Srivastava, J. M. Tripathi, S. R. Mohanty, & B. Panda, “Optimal over-
current relay coordination with distributed generation using hybrid particle
this study. The augmented ε-constraint method was used to swarm optimization–gravitational search algorithm.” Electr Pow Compo Sys,
compromise the total operation time of relays and the number vol. 44, no. 5, pp. 506-517, 2016.
of dual-setting DOCRs deployments. Different deployment [16] J. Radosavljević, and M. Jevtić. “Hybrid GSA-SQP algorithm for optimal
percentages of dual-setting DOCRs were optimally attained coordination of directional overcurrent relays.” IET Gener. Transm. Dis., vo.
10, no. 8, pp. 1928-1937, 2016.
for different coordination methods. It was deduced that the [17] Y. Serizawa, M. Myoujin, K. Kitamura, N. Sugaya, M. Hori, A.
three-point standard coordination approach leads to a higher Takeuchi, I. Shuto, M. Inukai. “Wide-area current differential backup
deployment percentage of dual-setting DOCRs against the protection employing broadband communications and time transfer
systems,” IEEE Trans. Power Del., vol. 13, no. 4, 1046-1052, 1998.
single-point approach (50% versus 41.67%). This notice [18] (2014, Jan. 15) [Online]. Available: http://www.easunreyrolle.com
portrays the necessity of more dual-setting DOCRs to attain /product.php?id=67B
more accurate coordination results. Non-standard coordination [19] (2014, Jan. 15) [Online]. Available: http://www.toshiba-tds.com
/tandd/products/pcsystems/en/grd100.htm
approach which relaxes characteristic coefficients in [20] H. M. Sharaf, H. H. Zeineldin, D. K. Ibrahim, and E. L. Essam, “A proposed
optimization process results in less deployment of dual-setting coordination strategy for meshed distribution systems with DG considering
DOCRs against the standard approach (33.33% versus user-defined characteristics of directional inverse time overcurrent
relays,” Int. J. Elec. Power., vol. 65, pp. 49-58, 2015.
41.67%); while, granting higher reduction in total operation [21] M. Shahidehpour, R. Allan, and P. Anderson. “Effect of protection systems on
time of relays. The non-standard approach, if to be applied in bulk power reliability evaluation,” IEEE Trans. Power Syst., vol. 9 no. 1 1994.
three-point manner, increased the deployment level of dual- [22] Terzija, Vladimir, et al., “Wide-area monitoring, protection, and control of
future electric power networks.” Proc. IEEE, Vol. 99, No .1, pp. 80-93, 2011.
setting DOCRs to assure a more reliable performance (41.67% [23] V. C. Nikolaidis, E. Papanikolaou, & A. S. Safigianni, “A
versus the 33.33% in single-point non-standard approach). communication-assisted overcurrent protection scheme for radial
Based on the simulation studies, the proposed methodology distribution systems with distributed generation”. IEEE Trans. Smart
Grid., vol. 7, no. 1, pp. 114-123, 2016.
can attain an effective and reliable coordination for fast fault
[24] C. J. Edwards, E. M. Davidson, S. D. McArthur, I. Watt, T. Cumming,
clearing purposes. “Flexible model-based alarm processing for protection performance
assessment and incident identification.” IEEE Trans. Power Syst., vol.
REFERENCES 28, no. 3, pp. 2584-2591, 2013.
[25] M. Dewadasa, A. Ghosh, and G. Ledwich, “Protection of distributed
[1] F. Guerin, D. Lefebvre, and A. B. Mboup, “Hybrid modeling for
generation connected networks with coordination of overcurrent relays,”
performance evaluation of multisource renewable energy systems,”
IEEE Trans. Autom. Sci. Eng., vol. 8, no. 3, pp. 570–580, Oct. 2011.

1551-3203 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2018.2821175, IEEE
Transactions on Industrial Informatics
10

IECON 2011-37th Annual Conference on IEEE Industrial Electronics Engineering Department, Urmia University, Urmia, Iran. His research
Society, 2011, pp. 924 929. interests are primarily centered in advanced power electronics and FACTS
[26] H. H. Zeineldin, H. M. Sharaf, D. K. Ibrahim, A. El-Zahab, and E. El-Din, applications in power systems as well as smart grid technologies.
“Optimal protection coordination for meshed distribution systems with DG
using dual setting directional over-current relays,” IEEE Trans. Smart Grid., Saeed Teimourzadeh (S’15) received the B.Sc.
vol. 6, no. 1, pp. 115-123, 2015. (Hons.) and M.Sc. (Hons.) degrees in electrical
[27] H. M. Sharaf, H. H. Zeineldin, and E. El-Saadany, “Protection coordination for engineering from the University of Tabriz, Tabriz,
microgrids with grid-connected and islanded capabilities using dual setting Iran, in 2012 and 2014, respectively. He is currently
directional overcurrent relays,” IEEE Trans. Smart Grid., to be published. pursuing the Ph.D. degree in the School of Electrical
[28] A. Arab, A. Khodaei, Z. Han, and S.K. Khator. "Proactive recovery of and Computer Engineering, University of Tehran,
electric power assets for resiliency enhancement." IEEE Access, vol. 3, pp. Tehran, Iran. Since 2016, he has been a Research
99-109, 2015. Associate with the Electrical and Computer
[29] G. Mavrotas, “Effective implementation of the e-constraint method in multi- Engineering Department at the Illinois Institute of
objective mathematical programming problems,” Appl. Math. Comput. Vol. Technology (IIT), Chicago, IL, USA. His research interests include microgrid
213, no. 2, pp. 455–465, 2009. protection, control and stability, and smart grid initiatives.
[30] Crow, Mariesa L. “Electric vehicle scheduling considering co-optimized
customer and system objectives.” IEEE Trans. Sustain. Energy, vol. 9, no. .1,
pp. 410-419, 2018. Farrokh Aminifar (SM’15) has been collaborating
[31] S. Agrawal, B. K. Panigrahi, and M. K. Tiwari, “Multiobjective particle swarm with the Robert W. Galvin Center for Electricity
algorithm with fuzzy clustering for electrical power dispatch,” IEEE Trans. Innovation with the Illinois Institute of Technology,
Evol. Comput., vol. 12, no. 5, pp. 529-541, 2008. Chicago, IL, USA, since March 2009. He is currently
[32] C. A. Rojas, J. R. Rodriguez, S. Kouro, & F. Villarroel, “Multiobjective fuzzy- an Assistant Professor with the School of Electrical
decision-making predictive torque control for an induction motor drive.” IEEE and Computer Engineering, University of Tehran,
Trans. Power Electron., vol. 32, no. 8, pp. 6245-6260, 2017. Tehran, Iran. His research interests include wide-area
[33] R. Asanga, K. H. Saman, C. W. Harry, “Self-Organizing hierarchical particle measurement systems, power system expansion
swarm optimizer with time-varying acceleration coefficients” IEEE Trans. planning and reliability assessment, and smart grid
Evol. Comput., vol. 8, no. 3, pp. 240–255, 2004. initiatives. Dr. Aminifar is serving the IEEE
[34] Y. Shen, G. Wang, C. Tao. “positive linear correlation particle swarm Transactions on Sustainable Energy and the IEEE Power Engineering Letters
optimization,” in RSKT, pp. 367-374, 2009, July. as the editor. He received the 2011 IEEE Iran Section Best Ph.D. Dissertation
[35] N. Watson and J. Arrillaga, Power systems electromagnetic transients Award, the 2013 IEEE/PSO Transactions Prize Paper Award, the 2015 IEEE
simulation. London, U.K.: IET, 2007, Ch.4. Iran Section Young Investigator Award, and the 2017 Outstanding Young
[36] IEEE Standard for Interconnecting Distributed Resources with Electric Scientist Award of Iran National Academy of Science.
Power Systems, IEEE Standard 1547-2003, pp. 1–28, 2003.
[37] R. Mohammadi, H. A. Abyaneh, H. M. Rudsari, S. H. Fathi, H. Rastegar,
“Overcurrent relays coordination considering the priority of
constraints” IEEE Trans. Power Del., vol. 26, no. 3, pp. 1927-1938, 2011.
[38] S. Teimourzadeh, F. Aminifar, M. Davarpanah, & J. M. Guerrero,
“Macroprotections for microgrids: Toward a new protection paradigm
subsequent to distributed energy resource integration” IEEE Ind. Electron
M, vol. 10, no. 3, pp. 6-18, 2016.

Amin Yazdaninejadi received the B.Sc. degree in


electrical engineering from Urmia University, Urmia,
Iran, in 2012, and the M.Sc. degrees from the electrical
engineering from Amirkabir University of Technology
(AUT), Tehran, Iran, in 2014. He is currently pursuing
the Ph.D. degree in the School of Electrical
Engineering, University of Urmia, Urmia, Iran. His
research interests include power system protection,
power system automation, and smart grid technologies

Sajjad Golshannavaz received the B.Sc. (Honors) and


M.Sc. (Honors) degrees in electrical engineering from
Urmia University, Urmia, Iran, in 2009 and 2011,
respectively. He received his Ph.D. degree in electrical
power engineering from School of Electrical and
Computer Engineering, University of Tehran, Tehran,
Iran, in 2015. Currently, he is an Assistant Professor in
Electrical Engineering Department, Urmia University,
Urmia, Iran. Since 2014 he has been collaborating with
the smart electric grid research laboratory, Department of Industrial
Engineering, University of Salerno, Salerno, Italy. His research interests are in
smart distribution grid operation and planning studies, design of distribution
management system (DMS), demand side management (DSM) concepts and
applications, microgrid design and operation studies, design of energy
management system (EMS), application of FACTS Controllers in Power
systems, application of intelligent controllers in power systems.

Daryoush Nazarpour received the B.Sc. degree from


the Iran University of Science and Technology,
Tehran, Iran, in 1982 and the M.Sc. and Ph.D. degrees
from University of Tabriz, Tabriz, Iran, in 1988 and
2005, respectively, all in electrical engineering. He is
currently an Associate Professor with the Electrical

1551-3203 (c) 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

You might also like