Computers and Electrical Engineering 74 (2019) 74–88

Contents lists available at ScienceDirect

Computers and Electrical Engineering

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

Overcurrent protection of AC microgrids using mixed

characteristic curves of relaysR
Mahamad Nabab Alam
Department of Electrical Engineering, Indian Institute of Technology, Roorkee 247667 India

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

Article history: Protection of AC microgrids with islanded and grid-connected modes of operation is a ma-
Received 27 May 2018 jor challenge as fault currents change drastically in the transition from one mode to the
Revised 17 December 2018
other. In this paper, optimum settings of directional overcurrent relays (DOCRs) consider-
Accepted 15 January 2019
ing different characteristic curves for the protection of AC microgrids with islanded and
grid-connected modes of operation is presented. In this work, all the relays are considered
Keywords: to be associated with three variables which are time multiplier-setting (TMS), plug-setting
AC microgrid (PS), and curve-setting (CS) to obtain their correct operating times. Here, TMS, PS, and CS
Curve selection are associated with time scaling, pickup current value and characteristic curve selection,
Directional overcurrent relays respectively, of a DOCR. The proposed protection coordination problem has been formu-
Genetic algorithm lated as a mixed-integer nonlinear programming problem and solved using the genetic
Protection coordination
algorithm in MATLAB environment. The effectiveness and suitability of the proposed ap-
Protection settings
proach have been demonstrated on 7 and 18 bus microgrids.
© 2019 Published by Elsevier Ltd.

1. Introduction

A microgrid is a group of interconnected loads and distributed energy resources (DERs) with the ability to operate in
islanded and grid-connected modes. It is an eco-friendly, reliable and convenient approach for utilizing DERs and integrating
them to the main grid. AC microgrids are very popular because of easy modeling, simple design and efficient performance.
However, protection of AC microgrid is a major challenge because the magnitude and direction of fault currents change
significantly in the islanded and grid-connected modes. Therefore, it is difficult to provide proper protection to the microgrid.
Protection issues of AC microgrids have been thoroughly discussed in [1].
For overcurrent protection of interconnected microgrids, directional overcurrent relays (DOCRs) are the efficient and eco-
nomical choice. In any protection scheme, the primary relay must initiate an operation to remove the faulted section quickly
to limit damage to the system. However, if the primary relay fails to remove the faulted section, then the corresponding
backup relay should operate. This feature of the protection scheme is the most desirable as the primary relay removes a
little portion of the system, whereas backup relay unnecessary removes a larger portion of the system [2]. A fixed mini-
mum time interval must be maintained between the operating times of the primary and the corresponding backup relay,
which is known as the coordination time interval (CTI). For ensuring the possibility of an unwanted power outage, proper
coordination among the DOCRs is necessary. Proper protection coordination can be achieved through optimum settings and
optimum curve selections of DOCRs.

R
Reviews processed and recommended for publication to the Editor-in-Chief by Guest Editor Dr. Pawan Kumar.
E-mail address: [email protected]

https://doi.org/10.1016/j.compeleceng.2019.01.003
0045-7906/© 2019 Published by Elsevier Ltd.
 M.N. Alam / Computers and Electrical Engineering 74 (2019) 74–88 75

The operating times of DOCRs are normally associated with their two parameters, namely time multiplier-setting (TMS)
and plug-setting (PS) [3]. These two parameters are continuous variables, which can have any value within their range in
numerical/digital type of relays. Moreover, these relays may follow different characteristic curves, such as inverse definite
minimum time (IDMT) or normal inverse (NI), very inverse (VI) or extremely inverse (EI), making the selection of TMS
and PS very difficult. A third parameter related to the selection of the characteristic curve of DOCR can be considered.
This parameter is termed as curve section or curve-setting (CS) of the relay. As the fault currents in the islanded mode of
microgrid are relatively lower, so DOCRs with VI or EI characteristic may be a good choice as it will take less time to operate
even for lower values (more than it’s PS value) of fault currents. The values of TMS, PS, and CS for all the relays which make
them coordinate properly, is called settings of the relay in this paper.
Protection coordination problem of DOCRs is formulated as a nonlinear programming (NLP) problem, and various opti-
mization techniques are used to solve it to obtain the optimum settings of the relays. In NLP formulation, the operating time
of a relay is a nonlinear function in TMS and PS of the relay. Further, TMS and PS are considered as the continuous vari-
able in NLP formulation. The protection coordination problem considering NLP formulation has been solved using various
analytical optimization methods in [4–6]. Subsequently, the application of various heuristic optimization methods has also
been proposed in the literature [7–10]. Mostly in the literature, only one fixed type of characteristic curve has been con-
sidered for all the relays to obtain the optimum protection coordination. However, it is to be noted that the consideration
of characteristic curve selection along with the optimum settings can improve the protection coordination because it will
provide more flexibility towards maintaining coordination constraints in microgrids. In the literature, the genetic algorithm
(GA) is proven to be a very efficient method for solving mixed-integer nonlinear programming (MINLP) problems related to
protection coordination of DOCRs [11].
Several schemes have been developed in the literature to address the issues related to overcurrent protection of AC
microgrids [12,13]. The works of the literature can be broadly classified into two categories:
1. Adaptive protection
2. Non-adaptive protection

In adaptive overcurrent protection of AC microgrids, settings of DOCRs are changed according to the system operating
conditions. Mostly, settings of the relays are changed in this scheme whenever a transition occurs between the islanded and
grid-connected modes. In [14], communication assisted voltage controlled DG integration has been studied for protection
of islanded microgrid by differentiating overcurrent and overloading. In [15], protection coordination of communication as-
sisted microprocessor-based relays for islanded microgrid has been discussed. For complete protection of microgrid using the
microprocessor and positive sequence component-based approach has been discussed by utilizing the phasor measurement
unit (PMU) with a digital communication system in [16]. Similarly, several studies have been performed in the literature
regarding adaptive protection of microgrids using DOCRs with communication system [17–19]. In the adaptive protection
schemes, communication channel plays a key role. Also, digital/numerical or microprocessor based DOCRs are needed. Al-
though, this scheme provides excellent protection coordination but very much reliant on the communication link. Also, this
scheme is very costly because of additional hardware requirements.
In non-adaptive overcurrent protection of AC microgrids, optimum setting of DOCRs are obtained with the ability to coor-
dinate properly in both islanded and grid-connected modes. In [20], protection coordination of microgrid with islanded and
grid-connected modes of operation has been discussed. The presented scheme does not require any communication link,
however, it requires programmable microprocessor-based DOCRs. In [21], laboratory-based protection coordination scheme
for islanded and grid-connected modes of microgrid has been investigated. In this scheme, several energy storage systems
including flywheel have been utilized in order to maintain smooth power supply. In [22], fault current limiters (FCLs) have
been utilized in order to obtain optimum settings of DOCRs suitable to protect microgrid with islanded and grid-connected
operating capabilities. Similarly, several other studies have been proposed with or without utilizing FCLs [23,24]. This scheme
is cheaper as it does not need any additional piece of equipment unlike in adaptive protection scheme. However, this scheme
is not very accurate because there are chances of miscoordination as well as the operating times of backup relays are rela-
tively higher and even sometimes backup relays do not operate at all. These are because of the robustness of the protection
scheme. However, the chances of failure are low.
Motivated with the above observations, optimum settings and curve selections of DOCRs for interconnected AC microgrid
is proposed in this paper. A third parameter is considered with each relay which is associated with the optimum selection
of the characteristic curve of the relay among NI, VI, and EI. The optimum protection coordination problem of AC microgrid
is formulated as a MINLP problem and solved using the GA in the MATLAB environment. Further, the robust protection
coordination setting of the microgrid is presented which can coordinate properly in islanded and grid-connected modes of
the microgrid. The feasibility of the developed algorithm has been investigated on a 7-bus AC microgrid derived from the
IEEE 14-bus test system. Further, the proposed approach has been validated on an 18-bus AC microgrid derived from the
IEEE 30-bus test system. Also, a comparison of the performance of the genetic algorithm with particle swarm optimization
(PSO) and differential evolution (DE) on the formulated problem has been presented.
The organization of the rest of this paper is as follows. In Section 2, issues related to microgrid protection are discussed.
In Section 3, the details of the coordination problem of DOCRs are described. In Section 4, relevant details about the pro-
posed optimization approach are discussed. Simulation results obtained by the proposed approach are included in Section 5.
Finally, the conclusions are presented in Section 6.
76 M.N. Alam / Computers and Electrical Engineering 74 (2019) 74–88

Fig. 1. Example of a typical microgrid.

2. Protection coordination issues in AC microgrids

Fig. 1 shows a typical microgrid. This microgrid is a 4-bus interconnected system supplied by inverter-based distributed
generation (IBDG) and a synchronous generator (SG). Also, the main grid is available to power the microgrid if required. The
IBDG gets power from photovoltaic (PV) panels and wind turbines (WTs) as DC power. The DC powers obtained from PV
panels and WTs are converted into AC power by using the individual inverter and then supplied to the microgrid at bus
4. An SG is present in the system to takes care of load fluctuations and the issues related to reactive power requirements.
During peak loading or extreme weather conditions, the main grid can supply the microgrid. Also, during excess power
generation from IBDG, the main grid can be supplied.
Commonly, microgrids can operate under the following two different scenarios:

1. Islanded mode
2. Grid-connected mode

In the islanded mode, the entire microgrid is supplied by the internal power generating sources. These internal sources
include IBDGs. The short-circuit currents are relatively smaller because of the power supply are through IBDGs in AC mi-
crogrids. In the grid-connected mode, the entire microgrid is supplied by the internal power generating sources as well as
the main external grid. In this mode, short-circuit currents are very high as compared to the islanded mode. Overcurrent
protection in both modes of microgrid may not be able to coordinate properly because of the following two reasons:

1. Change in the magnitude of fault currents

2. Change in the direction of fault currents

The optimum settings of DOCRs of islanded mode case may miscoordinate when used for protecting the microgrid in
the grid-connected mode and vice-versa. Therefore, two sets of optimum settings of DOCRs are required and also need to
change them as scenario changes to prevent any miscoordination which is a difficult task. In this paper, a common set of
the optimum settings of DOCRs has been proposed which is suitable to provide optimum protection coordination in both
islanded and grid-connected modes of microgrids.
It is to be noted that the values of fault currents are relatively lower in microgrids as compared to that in the main grid
which is mostly protected by using standard IDMT type characteristic curves of DOCRs. Therefore, adopting the same char-
acteristic curves for protection of microgrids will result in higher operating times of the relays [22–24]. So, the microgrids
may be forced to be under stress for a longer period during faults which may damage the entire system. As a result, VI
and EI characteristic curves along with standard IDMT characteristic curves for DOCRs will be more suitable for protection
of microgrids. So in this paper, the third parameter of the relay which is related to the selection of a suitable characteristic
curve for each relay is also considered. This third parameter of the relay is termed as CS which gives the information about
the type of characteristic curve followed by the relay.
Thus, protection coordination scheme of DOCRs used in protecting AC microgrid try to obtain the optimum values of TMS,
PS and CS for each relay in such a way that the obtained settings protect microgrids in both islanded and grid-connected
modes properly.
 M.N. Alam / Computers and Electrical Engineering 74 (2019) 74–88 77

Table 1
Characteristic curve coefficients of DOCRs.

Characteristic curve of relay λ η L CS

Very Inverse (VI) 13.5 1 0 1

Extremely Inverse (EI) 80.0 2 0 2
Normal Inverse (NI) 0.14 0.02 0 3

3. Problem formulation for overcurrent protection of AC microgrid

To obtain optimum coordination of relays protecting lines of the microgrid, optimum selection of characteristic curve and
settings of each relay are necessary. Thus, each relay is associated with three parameters which decide the operating time
of the relay. These three parameters are TMS, PS and CS. Curve setting of a relay gives the information about the type of
characteristic curve (among VI, EI, and NI) to be followed by that particular relay.
The optimum protection coordination problem of DOCRs is to minimize the sum of operating times of all the relays
corresponding to the maximum fault current as defined in [7,8] as,

m
CS
minimize top,i (1)
TMS,PS,CS
i=1

Subject to
CS CS
tob, j − top,i ≥ MCT (2)

TMSi,min ≤ TMSi ≤ TMSi,max (3)

PSi,min ≤ PSi ≤ PSi,max (4)

CS
ti,min ≤ top,i ≤ ti,max (5)

1 ≤ CS ≤ 3 (6)
In Eq. (1), top, i is the operating time of relay Ri and m is the total number of relays in the system. Further, superscript
CS ∈ {1, 2, 3} represents the type of characteristic curve of relay Ri . It is related to the selection of charcatetic curve for
various relays whose details are given in Table 1. In Eq. (2), top, i and tob, j are the operating time of primary relay Ri and
its backup relay Rj respectively for a fault at a line and MCT is the minimum coordination time (MCT) required. In Eq. (3),
TMSi,min and TMSi,max are the minimum and maximum limits on TMS, respectively, while in Eq. (4), PSi,min and PSi,max are
the same quantities on PS corresponding to relay Ri . In Eq. (5), ti,min and ti,max are the minimum and maximum operating
time of the relay Ri for the fault at any point in the zone of protection (i.e., lines).
The operating times of relays are obtained from their characteristic curves which are defined by IEC/IEEE [2] as,
λ × TMSi
top,i = +L (7)
(IF,i /PSi )η − 1
In Eq. (7) λ, η and L are the characteristic constants of the relays while IF, i is the fault current through the operating
coil of relay Ri . TMSi and PSi are the two variables of the relay Ri to be optimized. The characteristic coefficients of various
DOCRs are given in Table 1. Once the coordination constraint (i.e., Eq. (2)) is maintained at the maximum fault current then
at the lower values of fault currents, this constraint can automatically be maintained because of the inverse nature of the
characteristic curve of each relay.
It is to be noted that a third variable called CS is considered to be associated with each relay to be selected optimally. This
variable has an integer value of 1, 2 or 3 which decides the type of characteristic curve of the relay. Further, in this paper,
all the relays are considered to be digital/numerical relays which can accept continuous values of TMS and PS parameters.
However, the third parameter (which is CS) is also associated with each relay, accepts only integer values (1, 2 or 3). Thus,
the proposed protection coordination problem of AC microgrid becomes a MINLP problem. As the third parameters of each
relay are associated with the selection of curves and thus cannot be relaxed to continuous values, so it is difficult to apply
any analytical optimization techniques. The GA is having the ability to solve such MINLP problem efficiently. Therefore,
MATLAB based GA is adopted to solve the proposed protection coordination problem of AC microgrid in this work.
Various currents are scaled down using suitable current transformer (CT) for each relay as relay used to operate at very
low values of currents. The CT ratio (CTR) for ith relay is calculated as [2],
 
I f max,i
CTRi = max ILmax,i , (8)
20
78 M.N. Alam / Computers and Electrical Engineering 74 (2019) 74–88

Fig. 2. Flowchart of the proposed optimum settings.

In Eq. (8), Ifmax, i and ILmax, i are the values of maximum fault and load currents, respectively, for relay Ri .

4. Solution methodology of the proposed protection coordination problem

In this paper, the proposed optimum settings and curve selections problem of DOCRs has been solved using the GA in
MATLAB. Three sets of variables are used in the proposed formulation which are TMSs, PSs, and CSs. These three sets of
variables form the chromosome for GA.

Chromosome =| TMS| PS| CS|

where

TMS = [TMS1 , TMS2 , . . . , TMSm ]

PS = [PS1 , PS2 , . . . , PSm ]
CS = [CS1 , CS2 , . . . , CSm ].

In this paper, the limits on TMS and PS have considered in the in the range of [0.1, 1.1] and [0.5, 2.0] times of CT
secondary rating of each relay [8], respectively. Further, the limits on the operating times of relays are assumed in the range
of [0.1, 4] seconds [4]. Further, the MCT required for each primary-backup relay pairs is considered as 0.2 s [4,8].
It is to be noted that in this paper, backward-forward sweep method (BFSM) and bus impedance matrix (Zbus) approach
have been adopted for calculating the maximum load currents and various fault currents, respectively. Further, the various
fault has been applied to the middle of each line for ease of calculations. Also, for the maximum fault current calculation,
three-phase to the ground fault without any fault impedance has been considered [4,8]. It is to be noted that the directions
of the DOCRs are always fixed and are considered to be towards the line being protected.
The flowchart of the proposed approach for overcurrent protection of AC microgrid is shown in Fig. 2.
 M.N. Alam / Computers and Electrical Engineering 74 (2019) 74–88 79

Fig. 3. Single-line diagram of the 7-bus system.

5. Simulation results and discussion

The effectiveness and the suitability of the proposed protection coordination scheme of AC microgrid have been investi-
gated on a 7-bus test system. This test system is the low voltage (LV) section of the IEEE 14-bus system having eight lines
[6,25]. To completely protect this system, a total of 16 DOCRs are required (two relays on each line). Protection coordination
settings have been obtained under two scenarios and four functional characteristic curves of relays which are as follows:
Microgrid scenarios:

1. Islanded mode
2. Grid-connected mode

Relays functional characteristics:

1. Standard IDMT relays only

2. Very inverse relays only
3. Extremely inverse relays only
4. Mix of standard IDMT, very inverse and extremely inverse relays

Further, one more case (a common scenario) has been considered in which the effects of islanded and grid-connected
modes have been considered simultaneously for mixed characteristic curves of the relays.
Fig. 3 shows the 7-bus microgrid supplied by an SG connected at bus 1 (B1) and two IBDGs connected at buses 2 and
7. Further, this system in the grid-connected mode is supposed to be supplied through two buses (B3 and B6). Original
system information is available in [25]. The maximum generation limits on power suppliers at bus 1 is 50 MVA, at buses
2, and 7 is 20 MVA each and that at bus 3 and 6 is 60 MVA each. The maximum short-circuit level of generators at bus 1
is 250 MVA, at buses 2 and 7 is 80 MVA each whereas that of the external grid at bus 3 and 6 is 300 MVA each. All 16
DOCRs required to protect this system are shown against each line of the network in Fig. 3. This system is having a total of
22 primary-backup relay pairs among those 16 relays. Information about various primary-backup relay pairs and the fault
current passing through them for the 7-bus test system are given in Table 2. It is to be noted that three-phase fault has
been applied at the middle of each line to calculated various fault current given in this table.
Using Eq. (8), CTR for each relay of this test system is given in Table 3. It is to be noted that these CT ratios are kept
fixed in different scenarios of the microgrid.

5.1. Optimum settings for islanded mode

The optimum settings obtained for the 7-bus test system using various relay curves are given in Table 4 for the islanded
mode. The corresponding operating times of primary-backup relay pairs along with actual CTI for each pair are shown in
Fig. 4.
In case of the islanded mode of the 7-bus test system, from Table 4, it is observed that the sum of operating times of
all 16 relays considering VI, EI, NI and mixed of VI-EI-NI are 3.2457, 6.7142, 7.3148 and 2.0670 s respectively. The optimum
settings obtained using mixed characteristic curves of all relays, only VI and EI types of curves are present (last column in
80 M.N. Alam / Computers and Electrical Engineering 74 (2019) 74–88

Table 2
Primary-backup relay pairs and various fault currents in the 7-bus microgrid.

Faulty Serial Primary Backup Islanded mode Grid-connected mode

Line No. Relay Relay Ifprim (A) Ifback (A) Ifprim (A) Ifback (A)

L1 1 1 4 3612 384 4830 1914

2 1 6 3612 564 4830 812
3 2 8 2305 726 3435 2069
L2 4 3 2 4844 959 5737 1641
5 3 6 4844 333 5737 694
6 4 13 2140 954 6430 1587
7 4 16 2140 1057 6430 543
L3 8 5 2 3968 928 5503 1308
9 5 4 3968 323 5503 1600
10 6 15 2345 835 3357 1980
L4 11 7 1 3087 1469 3519 1988
12 8 10 1963 1403 5374 1170
L5 13 9 7 2678 2087 8019 1948
14 10 12 2076 1886 2792 2607
L6 15 11 9 2197 2008 4777 4648
16 12 14 2470 2419 3590 3544
L7 17 13 11 1548 1486 2926 2879
18 14 3 3862 2548 6626 2206
19 14 16 3862 1173 6626 964
L8 20 15 3 3048 2091 5673 1219
21 15 13 3048 861 5673 1245
22 16 5 2606 1233 3115 1829

Table 3
CT ratio for each relay in the 7-bus mi-
crogrid.

Relay CTR Relay CTR

1 20 0 0/5 9 2500/5
2 10 0 0/5 10 1200/5
3 30 0 0/5 11 1200/5
4 20 0 0/5 12 2500/5
5 1600/5 13 800/5
6 10 0 0/5 14 30 0 0/5
7 2500/5 15 1600/5
8 1600/5 16 1600/5

Table 4
Optimum settings of relays for islanded mode of operation in the 7-bus microgrid.

Relays VI EI NI VI-EI-NI

TMS (s) PS (A) TMS (s) PS (A) TMS (s) PS (A) TMS (s) PS (A) CS

1 0.2386 0.5226 0.3859 0.8901 0.1003 1.6211 0.2297 0.6643 2

2 0.2079 0.6414 0.2294 1.1720 0.1313 1.4489 0.1487 1.0523 2
3 0.1630 0.7728 0.4969 0.6762 0.2197 0.5521 0.2010 0.6347 2
4 0.10 0 0 0.5001 0.1416 0.5004 0.1062 0.5031 0.10 0 0 0.50 0 0 1
5 0.1639 0.5780 0.5159 0.9691 0.2012 0.5449 0.3559 0.7336 2
6 0.1493 0.5544 0.5453 0.5607 0.1481 0.5192 0.1390 0.5934 1
7 0.2758 0.5082 0.7615 0.7599 0.1988 0.7613 0.1488 0.7475 2
8 0.1002 0.5003 0.3063 0.7059 0.10 0 0 0.7619 0.1543 0.5499 2
9 0.1010 1.0071 0.2782 1.0508 0.1710 0.7655 0.2031 0.5816 2
10 0.10 0 0 1.1218 0.4215 0.9334 0.1884 0.5137 0.3445 0.6071 2
11 0.1563 0.9893 0.4637 0.9996 0.1714 0.9165 0.1395 1.0458 2
12 0.1896 0.5117 0.5818 0.6596 0.1767 0.5900 0.1084 0.7294 2
13 0.2264 0.5963 0.5490 1.2571 0.1552 0.7444 0.1794 0.9067 2
14 0.1817 0.6783 0.2580 1.0154 0.1216 1.3517 0.20 0 0 0.6249 2
15 0.1053 0.6261 0.1993 0.8571 0.1201 0.5358 0.1597 0.8393 2
16 0.1251 0.8234 0.2238 0.9123 0.1914 0.6155 0.1658 0.7045 2
 top 3.2457 s 6.7142 s 7.3148 s 2.0670 s
 M.N. Alam / Computers and Electrical Engineering 74 (2019) 74–88 81

Fig. 4. Optimum values of operating times of primary-backup relay pairs in islanded mode of microgrid.

Table 4). Further, from Table 4, it is observed that the optimum settings obtained corresponding to the mixed (VI-EI-NI)
characteristic curves of relays are the best as it takes the least sum of operating time of all 16 DOCRs in the islanded mode
of the 7-bus system. Further, from Fig. 4, it can be observed that CTI is always more than the MCT fixed.

5.2. Optimum settings for grid-connected mode

The optimum settings obtained for the 7-bus test system using various relay curves are given in Table 5 for the grid-
connected mode. The corresponding operating times of primary-backup relay pairs along with actual CTI for each pair are
shown in Fig. 5.
In case of the grid-connected mode of the 7-bus test system, from Table 5, it is observed that the sum of operating
times of all 16 relays considering VI, EI, NI and mixed of VI-EI-NI are 2.4392, 1.6681, 7.2041 and 1.6684 s respectively. As
in the previous case, the optimum settings obtained using mixed characteristic curves of all relays, in this case also, only
VI and EI types of curves are present as indicated in the last column in Table 5. Further, from Table 5 it is observed that
the optimum settings obtained corresponding to EI and mixed characteristic curves (VI-EI-NI) are better than the rest with
1.6681 and 1.6684 s, respectively. However, the results corresponding to the mixed characteristic curves are better than
those to EI curves. The reason for this lies in the higher values of the operating times of backup relays with the settings
corresponding to EI curves as compared to those of VI-EI-NI curves which can clearly be observed from Fig. 5 (Fig. 5(b) and
5(d), respectively).

5.3. Proposed common optimum settings

In the proposed common optimum settings of DOCRs for overcurrent protection of AC microgrid, the effects of both is-
landed and grid-connected modes of microgrid are combined before applying the optimization algorithm. Thus, the total
82 M.N. Alam / Computers and Electrical Engineering 74 (2019) 74–88

Table 5
Optimum settings of relays for grid-connected mode of operation in the 7-bus microgrid.

Relays VI EI NI VI-EI-NI

TMS (s) PS (A) TMS (s) PS (A) TMS (s) PS (A) TMS (s) PS (A) CS

1 0.1628 0.7951 0.2380 0.8727 0.2237 0.6350 0.2632 0.8302 2

2 0.2422 0.7589 0.4984 0.8591 0.2104 0.7894 0.3367 1.0446 2
3 0.10 0 0 0.8553 0.1391 0.9023 0.1419 0.9173 0.2681 0.6513 2
4 0.1397 0.8093 0.4976 0.8047 0.1582 1.0445 0.4214 0.8742 2
5 0.1512 0.8031 0.2447 1.2261 0.2171 0.5177 0.3187 1.0749 2
6 0.1704 0.6992 0.4630 0.8710 0.1133 1.3591 0.4318 0.9017 2
7 0.1534 0.6704 0.2388 0.5079 0.1495 1.0693 0.1043 0.7660 2
8 0.1553 0.9289 0.3974 0.9404 0.1617 1.2100 0.3357 1.0229 2
9 0.3805 0.6870 0.4568 0.8513 0.2367 1.0822 0.2674 1.1093 2
10 0.1066 0.8942 0.2977 0.7522 0.1098 1.4308 0.4374 0.6210 2
11 0.3527 0.8387 0.4242 1.1239 0.2284 1.2014 0.4202 1.1292 2
12 0.1678 0.6460 0.1266 0.8850 0.1572 0.7586 0.1019 0.9829 2
13 0.2892 0.7089 0.2368 1.3251 0.2267 0.7369 0.4092 1.0093 2
14 0.1681 0.9307 0.1019 1.2155 0.1843 0.9154 0.5369 0.5357 2
15 0.1984 0.6380 0.5505 0.8438 0.1568 0.7462 0.5674 0.8311 2
16 0.1413 0.5323 0.4664 0.5033 0.1792 0.5252 0.1038 0.6483 1
 top 2.4392 s 1.6681 s 7.2041 s 1.6684 s

Fig. 5. Optimum values of operating times of primary-backup relay pairs in grid-connected mode of microgrid.
 M.N. Alam / Computers and Electrical Engineering 74 (2019) 74–88 83

Fig. 6. Flowchart of the proposed common optimum settings.

Table 6
Common optimum settings of relays for islanded and grid-connected opera-
tions in the 7-bus microgrid.

Relays TMS (s) PS (A) CS Relays TMS (s) PS (A) CS

1 0.4541 0.6326 2 9 0.4061 0.5045 1

2 0.3633 1.0057 2 10 0.1996 1.0378 2
3 0.1238 0.9960 1 11 0.4552 1.1122 2
4 0.10 0 0 0.5012 3 12 0.1106 0.9867 2
5 0.4281 0.9279 2 13 0.3689 1.1405 2
6 0.1629 0.7302 1 14 0.4411 0.6023 2
7 0.1661 0.9772 2 15 0.2356 0.5405 1
8 0.1646 0.7702 1 16 0.1673 0.6778 1
 top 2.4392 s

number of relays remain the same whereas the total number of primary-backup relays pairs becomes double. As the coor-
dination constraints of islanded and grid-connected modes are considered together, so the obtained settings can coordinate
properly in both modes of the microgrid. Overall steps followed to obtain the common optimum settings are given in the
flowchart shown in Fig. 6.
The common optimum settings of relays which can coordinate properly in both islanded and grid-connected modes of
microgrid are given in Table 6 and the corresponding operating times of primary-backup relay pairs are shown in Fig. 7.
It is to be noted that in common optimum settings coordination constraints of both islanded and grid-connected modes of
microgrid have been considered simultaneous. In Fig. 7, the first 22 pairs of primary-backup relay pairs are corresponding
to the grid-connected mode whereas the remaining 22 pairs are corresponding to the islanded mode of the microgrid.
From Table 6, it is observed that the optimum settings obtained corresponding to the mixed (VI-EI-IN) characteristic
curves of relays takes a total 2.4392 s by all 16 DOCRs in the common mode of the 7-bus microgrid. In this case, the
optimum settings obtained using mixed characteristic curves of all relays and all the three characteristic curves (VI, EI, and
NI) are presents as given in Table 6. Also, from Fig. 7, it can be observed that the operating times and CTI of primary-backup
relay pairs are less than those corresponding to islanded (Fig. 4(d)) and grid-connected (Fig. 5(d)) modes of microgrid.
Thus, the common optimum settings of DOCRs are the best one to protect AC microgrids in both islanded and grid-
connected modes.
84 M.N. Alam / Computers and Electrical Engineering 74 (2019) 74–88

Fig. 7. Optimum values of operating times of primary-backup relay pairs in common mode (islanded and grid-connected) of the 7-bus microgrid.

Table 7
Cases of coordination constraint violation in the 7-bus microgrid

Settings Characteristic curve Number of coordination constraint violation

Calculated considered Islanded mode Grid connected mode

Islanded VI 0 14
EI 0 5
NI 0 13
VI-EI-NI 0 12
Grid-connected VI 7 0
EI 4 0
NI 8 0
VI-EI-NI 6 0
Common settings VI-EI-NI 0 0

5.4. Analysis of coordination constraint violation

The optimum protection coordination settings obtained under various microgrid scenarios and considering different char-
acteristic curves are analyzed by counting the cases of coordination constraint violation (i.e., Eq. (2)) with the settings of
one scenario to the other one. A coordination constraint is said to be violated for a given fault if any one of the following
conditions is observed:

1. Backup relay of a primary-backup relay pair operate before its primary relay
2. Actual CTI becomes less than MCT fixed for a primary-backup relay pair
3. Backup relay of a primary-backup relay pair takes longer time or remain idle

Table 7 gives the details about the number of coordination constraint violation in the 7-bus microgrid.
From Table 7, it is observed that there are many coordination constraint violations when optimum setting obtained for
islanded mode is applied to protect grid-connected mode of the microgrid. Similarly, many coordination constraint violations
can be observed when optimum settings obtained for grid-connected mode is applied to protect the islanded mode of
the microgrid. On the other hand, the optimum settings obtained considering the common optimum settings result in no
constraint violation in protecting both islanded and grid-connected modes of the microgrid.
Therefore, the optimum settings of relays obtained for the common mode are robust enough to coordinate properly in
both islanded and grid-connected modes of microgrids.

5.5. Validation of the proposed protection scheme on a larger microgrid

To validate the proposed protection scheme on a larger microgrid test system the following have been observed. An 18-
bus microgrid has been derived from the IEEE 30-bus power system. This test microgrid is the LV section (33 kV) of the IEEE
30-bus system having 22 lines. Fig. 3 shows the 18-bus microgrid supplied by an SG connected at bus 1 and three IBDGs
connected at buses 4, 11 and 18. This system can be grid-connected through three buses (B1, B2 and B16) as shown in the
figure. Original system information is available in [25]. The maximum generation limits on SG is 50 MVA and on IBDG is 20
MVA each. The maximum short-circuit level on the SG is 250 MVA and on each IBDG is 80 MVA. The maximum short-circuit
contribution from the main grid through each point of common coupling (external grid point) is 300 MVA.
 M.N. Alam / Computers and Electrical Engineering 74 (2019) 74–88 85

Fig. 8. Single-line diagram of the 18-bus microgrid.

To completely protect this system, a total of 44 DOCRs are required (two relays on each line) as shown in Fig. 8. This
system there are 72 primary-backup relay pairs among those 22 relays. These pairs are given in Table 8. To validate the
proposed protection scheme, only the common scenario (as discussed in Section 5.3 for the 7-bus microgrid) has been
considered for this microgrid. Both islanded and grid-connected modes have been considered simultaneously along with
the mixed characteristic curves of the relays in this test system. There are some primary-backup relay pairs in which the
fault current passing through the corresponding backup relays are small. As a result, the operating times of the backup
relays of these pairs become large. Such pairs can be identified and removed from the coordination constraint during the
optimization as the coordination constraint in these cases will always be satisfied.
The common optimum settings of relays obtained using GA for the 18-bus microgrid are given in Table 9 and the cor-
responding operating times of primary-backup relay pairs are shown in Fig. 9. It is to be noted that in common optimum
settings coordination constraints of both islanded and grid-connected modes of microgrid have been considered simultane-
ous. In Fig. 9, the first 72 pairs of primary-backup relay pairs are corresponding to the grid-connected mode whereas the
remaining pairs are corresponding to the islanded mode of the microgrid [6].
From Table 9, it is observed that the optimum settings obtained using GA corresponding to the mixed (VI-EI-IN) charac-
teristic curves of relays take a total 11.2509 s by all 44 DOCRs. Also, from Fig. 9, it can be observed that all the primary-
backup relay pairs satisfy the coordination constraint requirement. Thus, the proposed approach also provides the optimum
settings of DOCRs for this 18-bus system which can protect this AC microgrid in both islanded and grid-connected modes.
This test system indicates that the proposed scheme can be applied even on AC microgrids of larger size. In the next sec-
tion, a comparative study of the common optimum settings obtained by GA has been dome with those obtained by two
well-explored metaheuristic algorithm [10].

5.6. Comparison of the results obtained with some other optimization techniques

To show the effectiveness of the considered GA, a comparative study has been performed with PSO and DE [10]. Pop-
ulation size in all these three techniques has been considered 100. The other parameters of these algorithms have been
86 M.N. Alam / Computers and Electrical Engineering 74 (2019) 74–88

Table 8
Primary-backup relay pairs and various fault currents in the 18-bus microgrid.

Faulty Serial Primary Backup Faulty Serial Primary Backup

Line No. Relay Relay Line No. Relay Relay

L1 1 1 4 L10 37 19 11
2 1 6 38 19 15
3 1 8 39 19 18
4 2 21 40 20 32
L2 5 3 2 L11 41 21 13
6 3 6 42 22 1
7 3 8 L12 43 23 17
8 4 25 44 24 26
L3 9 5 2 L13 45 25 23
10 5 4 46 26 3
11 5 8 L14 47 27 5
12 6 28 48 28 7
L4 13 7 2 49 28 30
14 7 4 L15 50 29 7
15 7 6 51 29 27
16 8 27 52 30 31
17 8 30 53 30 34
L5 18 9 12 L16 54 31 19
19 9 14 55 32 29
20 10 16 56 32 34
L6 21 11 10 L17 57 33 29
22 11 14 58 33 31
23 12 15 59 34 38
24 12 18 L18 60 35 33
25 12 20 61 35 38
L7 26 13 10 L19 62 37 33
27 13 12 63 38 40
28 14 22 64 38 42
L8 29 15 9 L20 65 39 37
30 16 11 66 39 42
31 16 18 67 40 44
32 16 20 L21 68 41 37
L9 33 17 11 69 41 40
34 17 15 70 42 43
35 17 20 L22 71 43 39
36 18 24 72 44 41

Fig. 9. Optimum values of operating times of primary-backup relay pairs in common mode (islanded and grid-connected) of the 18-bus microgrid.
 M.N. Alam / Computers and Electrical Engineering 74 (2019) 74–88 87

Table 9
Common optimum settings of relays for islanded and grid-connected operations
in the 18-bus microgrid.

Relays TMS (sec) PS (A) CS Relays TMS (sec) PS (A) CS

1 0.5242 1.0635 2 23 0.6598 1.0055 1

2 0.5382 0.8449 2 24 0.6304 1.3784 2
3 0.41 1.7326 2 25 0.4225 1.5257 2
4 0.5658 0.9686 2 26 0.4828 1.4625 2
5 0.2899 1.2957 2 27 0.3548 1.0414 2
6 0.2195 1.0682 1 28 0.5479 0.9787 2
7 0.5383 0.86 2 29 0.2497 1.1326 2
8 0.5668 0.753 2 30 0.3747 1.3403 2
9 0.9138 0.8207 2 31 0.6164 1.3925 2
10 0.623 0.9166 2 32 0.5799 1.0723 2
11 0.3439 0.869 1 33 0.6192 1.5099 2
12 0.7288 0.9331 2 34 0.7826 0.8528 2
13 0.2945 1.2567 2 35 0.7677 0.9989 2
14 0.4978 1.4679 2 36 0.7407 1.164 2
15 0.5423 1.1825 2 37 0.5125 1.698 2
16 0.9186 0.7064 2 38 0.2204 0.9695 1
17 0.5243 1.4744 2 39 0.6585 0.8893 2
18 0.2308 1.6815 2 40 0.7436 1.3474 2
19 0.4559 1.4358 2 41 0.3442 1.2026 2
20 0.3662 1.1982 2 42 0.6515 1.1534 3
21 0.4699 1.0859 2 43 0.2963 1.554 2
22 0.7036 1.2 2 44 0.7434 1.0273 2
 top 11.2509 s

Table 10
Summary of results obtained for common optimum settings for the 18-bus microgrid after 100
independent runs.

Microgrids Methods Objective function value (s) Standard Average elapsed

Best Mean Worst deviation time (s)

7-bus GA 2.2373 4.795 10.7683 2.6532 96.7315

PSO 4.586 5.6215 6.7676 0.8435 87.0094
DE 3.4818 4.098 4.5676 0.3585 39.2953
18-bus GA 11.2509 12.7435 14.1262 0.8454 53.725
PSO 14.3181 18.0809 20.4978 2.1996 166.4672
DE 16.8837 19.2106 23.0314 2.2332 75.0359

considered as the standard parameters mentioned in [10]. Each algorithm has been executed 100 times and statistics of the
obtained results have been presented in Table 10.
From Table 10, it can be observed that the optimum settings obtained using GA are better than PSO and DE for both the
test systems (7 and 18 bus microgrids). The best values of the objective function (i.e., the sum of operating times of all the
relays) for the 7 and 18 bus microgrids obtained using GA are always less than those obtained using PSO and DE.
This analysis demonstrates the superiority of GA in solving protection coordination problem considering mixed-
characteristics curves of DOCRs. Hence, the proposed formulation and GA can be used to find the best possible characteristic
curves and settings of DOCRs for protection of AC microgrids.

6. Conclusion

The optimum protection coordination for AC microgrids using directional overcurrent relays has been analyzed. The pro-
posed approach considers optimum characteristic curve selection and optimum settings of the relay parameters which coor-
dinate properly in both islanded and grid-connected modes of microgrids. For this, protection coordination of AC microgrid
has been formulated as a mixed-integer nonlinear programming problem subject to maintaining proper coordination among
the relays. To solve this problem genetic algorithm has been adopted in the MATLAB environment. The effectiveness and
suitability of the developed approach have been tested on a 7-bus microgrid derived from the IEEE 14-bus system and an
18-bus microgrid derived from the IEEE 30-bus system. On the basis of the test results, it has been found that the optimum
settings obtained for the islanded mode of microgrid, coordinates properly in the islanded mode but miscoordinates in the
grid-connected mode and the optimum settings obtained for the grid-connected mode of microgrid, coordinates properly in
the grid-connected mode but miscoordinates in the islanded mode. On the other hand, it has been observed that the opti-
mum settings obtained considering the fault currents of the islanded and grid-connected modes of microgrid simultaneously
give robust settings of the relays which coordinate properly in both islanded and grid-connected modes of the microgrid.
88 M.N. Alam / Computers and Electrical Engineering 74 (2019) 74–88

Also, it has been found that the overcurrent protection of AC microgrid using mixed characteristic curves of relays coordi-
nates better than considering a single characteristic curve of all the relays.

References

[1] Memon AA, Kauhaniemi K. A critical review of AC microgrid protection issues and available solutions. Electr Power Syst Res 2015;129:23–31.
[2] Gers J, Holmes E. Protection of electricity distribution networks. IET power and energy series. 3rd edition. Institution of Engineering and Technology;
2011. 9781849192231
[3] Shih MY, Salazar CAC, Enrquez AC. Adaptive directional overcurrent relay coordination using ant colony optimisation. IET Generat Trans Distribut
2015;9(14):2040–9. doi:10.1049/iet-gtd.2015.0394.
[4] Alam MN, Das B, Pant V. An interior point method based protection coordination scheme for directional overcurrent relays in meshed networks. Int J
Electr Power Energy Syst 2016;81:153–64.
[5] Alkaran DS, Vatani MR, Sanjari MJ, Gharehpetian GB, Yatim AH. Overcurrent relays coordination in interconnected networks using accurate analytical
method and based on determination of fault critical point. IEEE Trans Power Delivery 2015;30(2):870–7. doi:10.1109/TPWRD.2014.2330767.
[6] Alam MN. Adaptive protection coordination scheme using numerical directional overcurrent relays. IEEE Trans Ind Inf 2018. 1–1. doi:10.1109/TII.2018.
2834474.
[7] Mansour MM, Mekhamer S, El-Kharbawe N-S. A modified particle swarm optimizer for the coordination of directional overcurrent relays. IEEE Trans
Power Delivery 20 07;22(3):140 0–10. doi:10.1109/TPWRD.2007.899259.
[8] Bedekar P, Bhide S. Optimum coordination of directional overcurrent relays using the hybrid GA-NLP approach. IEEE Trans Power Delivery
2011;26(1):109–19. doi:10.1109/TPWRD.2010.2080289.
[9] Amraee T. Coordination of directional overcurrent relays using seeker algorithm. IEEE Trans Power Delivery 2012;27(3):1415–22. doi:10.1109/TPWRD.
2012.2190107.
[10] Alam MN, Das B, Pant V. A comparative study of metaheuristic optimization approaches for directional overcurrent relays coordination. Electr Power
Syst Res 2015;128:39–52.
[11] Alkaran DS, Vatani MR, Sanjari MJ, Gharehpetian GB, Naderi MS. Optimal overcurrent relay coordination in interconnected networks by using fuzzy-
based GA method. IEEE Trans Smart Grid 2016;PP(99). 1–1. doi:10.1109/TSG.2016.2626393.
[12] Hooshyar A, Iravani R. A new directional element for microgrid protection. IEEE Trans Smart Grid 2017;PP(99). 1–1. doi:10.1109/TSG.2017.2727400.
[13] Nikkhajoei H, Lasseter RH. Microgrid protection. In: Proceedings of the IEEE power engineering society general meeting; 2007. p. 1–6. doi:10.1109/PES.
2007.385805.
[14] Etemadi AH, Iravani R. Overcurrent and overload protection of directly voltage-controlled distributed resources in a microgrid. IEEE Trans Ind Electron
2013;60(12):5629–38. doi:10.1109/TIE.2012.2229680.
[15] Lai K, Illindala MS, Haj-ahmed MA. Comprehensive protection strategy for an islanded microgrid using intelligent relays. IEEE Trans Ind Appl
2017;53(1):47–55. doi:10.1109/TIA.2016.2604203.
[16] Mirsaeidi S, Said DM, Mustafa MW, Habibuddin MH. A protection strategy for micro-grids based on positive-sequence component. IET Renewable
Power Gener 2015;9(6):600–9. doi:10.1049/iet-rpg.2014.0255.
[17] Zarei SF, Parniani M. A comprehensive digital protection scheme for low-voltage microgrids with inverter-based and conventional distributed genera-
tions. IEEE Trans Power Delivery 2017;32(1):441–52. doi:10.1109/TPWRD.2016.2566264.
[18] Cintuglu MH, Ma T, Mohammed OA. Protection of autonomous microgrids using agent-based distributed communication. IEEE Trans Power Delivery
2017;32(1):351–60. doi:10.1109/TPWRD.2016.2551368.
[19] Habib HF, Lashway CR, Mohammed OA. A review of communication failure impacts on adaptive microgrid protection schemes and the use of energy
storage as a contingency. IEEE Trans Ind Appl 2017;PP(99). 1–1. doi:10.1109/TIA.2017.2776858.
[20] Zamani MA, Sidhu TS, Yazdani A. A protection strategy and microprocessor-based relay for low-voltage microgrids. IEEE Trans Power Delivery
2011;26(3):1873–83. doi:10.1109/TPWRD.2011.2120628.
[21] Zhao B, Zhang X, Chen J. Integrated microgrid laboratory system. IEEE Trans Power Syst 2012;27(4):2175–85. doi:10.1109/TPWRS.2012.2192140.
[22] Najy WKA, Zeineldin HH, Woon WL. Optimal protection coordination for microgrids with grid-connected and islanded capability. IEEE Trans Ind
Electron 2013;60(4):1668–77. doi:10.1109/TIE.2012.2192893.
[23] Dehghanpour E, Karegar H, Kheirollahi R, Soleymani T. Optimal coordination of directional overcurrent relays in microgrids by using cuckoo-linear
optimization algorithm and fault current limiter. IEEE Trans Smart Grid 2016;PP(99). 1–1. doi:10.1109/TSG.2016.2587725.
[24] Sharma A, Panigrahi BK. Phase fault protection scheme for reliable operation of microgrids. IEEE Trans Ind Appl 2017;PP(99) . 1–1. doi:10.1109/TIA.
2017.2787691.
[25] Christie R. Power system test cases; 1993. www.ee.washington.edu/research/pstca.

Mahamad Nabab Alam received the Ph.D. degree in electrical engineering from Indian Institute of Technology, Roorkee, India, in 2018. His current research
interests include active distribution system protection, economic operation of power systems, and networked microgrids.