This article has been accepted for inclusion in a future issue of this journal.

Content is final as presented, with the exception of pagination.

IEEE TRANSACTIONS ON POWER DELIVERY 1
Optimal Coordination of Directional Overcurrent
Relays Using a New TimeCurrentVoltage
Characteristic
Khaled A. Saleh, Student Member, IEEE, H. H. Zeineldin, Senior Member, IEEE, A. Al-Hinai, Senior Member, IEEE,
and Ehab F. El-Saadany, Senior Member, IEEE
AbstractWith the integration of distributed generation (DG) to
meshed distribution systems, the operating time of the protective
system becomes a major concern in order to avoid nuisance DG
tripping. This paper proposes a newtimecurrentvoltage tripping
characteristic for directional overcurrent relays (DOCRs) that can
achieve a higher possible reduction of overall relays operating time
in meshed distribution networks. The proposed tripping charac-
teristic is described in detail. Moreover, the protection coordina-
tion problem is formulated as a constrained nonlinear program-
ming problem to determine the optimal relay settings. The pro-
posed characteristic is tested on the power distribution system of
the IEEE 14 bus and IEEE 30 bus with inverter-based and syn-
chronous-based DG units. The outcome of this study reveals that
the newtripping characteristic for DOCRs achieves notable reduc-
tion in total relays operating time over the conventional charac-
teristic.
Index TermsDirectional overcurrent relay (DOCR), dis-
tributed generation (DG), protection coordination, tripping
characteristic.
I. INTRODUCTION
P
ROTECTION coordination is a process of determining
the primary protective devices responsible for clearing the
fault as quickly as possible for each fault location, taking into ac-
count that in the event that any of the devices fail, each should be
backed up by another protective device [1]. The introduction of
distributed generation (DG) to the distribution network and the
recent drive toward smart grids is expected to transform the dis-
tribution networks into a meshed structure resembling intercon-
nected subtransmission systems [2]. The occurrence of faults on
Manuscript received July 01, 2013; revised September 26, 2013, December
18, 2013, and February 25, 2014; accepted July 17, 2014. This work was sup-
ported by the Masdar Institute of Science and Technology. Paper no. TPWRD-
00745-2013.
K. A. Saleh and A. Al-Hinai are with the iEnergy and the Department of
Electrical Engineering and Computer Science, Masdar Institute of Science and
Technology, Abu Dhabi, United Arab Emirates (e-mail: [email protected]).
H. H. Zeineldin is with the Faculty of Engineering, Cairo University, Giza,
Egypt and also with the iEnergy and the Department of Electrical Engineering
and Computer Science, Masdar Institute of Science and Technology, Abu Dhabi,
United Arab Emirates.
E. F. El-Saadany is with the Department of Electrical and Computer En-
gineering, University of Waterloo, Waterloo, ON N2L 3G1 Canada (e-mail:
[email protected]).
Color versions of one or more of the gures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identier 10.1109/TPWRD.2014.2341666
such meshed systems will cause bidirectional ow of short-cir-
cuit currents. Moreover, the presence of DG will affect the mag-
nitudes and directions of the short-circuit currents resulting in a
possible failure of the coordination between overcurrent relays
[3]. Hence, protection schemes based on directional overcurrent
relays (DOCRs) are proposed for such systems [4][7].
One of the most popular solutions proposed in the litera-
ture to solve the protection coordination problem of intercon-
nected systems is the use of optimization methods [4][17].
The main objective is to achieve minimum possible tripping
times by obtaining the optimal settings of each DOCR. In [9],
an MINLP formulation for the protection coordination problem
is proposed. Yet, in [10][13], the problem is linearized and
formulated as a linear-programming (LP) problem. More re-
cently, heuristic techniques, such as particle swarmoptimization
(PSO), genetic algorithm (GA), and evolutionary methods have
been proposed to solve the complex and nonconvex DOCRs
protection coordination problem [14][17]. Finally, in [5][7]
and [18], the protection coordination problemis relaxed and for-
mulated as nonlinear programing (NLP) problem.
Usually, DOCRs operate based on the conventional inverse
timecurrent characteristic. According to [19], further reduc-
tion of overall tripping time is possible by investigating the
tripping time characteristic equation. Since in todays tech-
nology OCRs/DOCRs are microprocessor based, the tripping
time characteristic equation can be changed by reprogramming
the relay. In [19], a new tripping characteristic equation was
proposed for OCR based on the logarithmic function. The
proposed logarithmic characteristic replaces the conventional
inverse timecurrent characteristic equation of OCRs. The
results indicate that the use of the new characteristic equation
leads to a noticeable reduction of the tripping time with respect
to the conventional characteristic for radial systems using OCR
and, more specically, with inverter-based DG [19]. In this
paper, both inverter-based and synchronous-based DG units are
considered. Synchronous generator-based DG units are taken
into account due to the fact that their fault current contribution
is much higher than inverter-based DG units and, thus, they
have a much more profound effect on protection coordination
[20]. On the other hand, although inverter-based DG units fault
current contribution is only in the range of 1.5-2 p.u, they are
more dominant than synchronous-based DG units especially
with increased renewable energy integration [20].
0885-8977 2014 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 inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
2 IEEE TRANSACTIONS ON POWER DELIVERY
This paper proposes a new nonstandard relay characteristic
that depends not only on current magnitude but also on voltage
magnitude for determining the relay operating time. The pro-
posed characteristic, which will be denoted as the timecur-
rent-voltage characteristic, can be implemented in micropro-
cessor-based DOCRs. The proposed relay characteristic should
achieve minimum overall tripping time in the protection co-
ordination problem of meshed distribution networks equipped
with directly connected synchronous generator-based DG units
as well as inverter-based DGunits. In order to test and verify the
effectiveness of the proposed approach, the power distribution
system of the IEEE 14-bus and IEEE 30-bus standard test sys-
tems are modeled with the introduction of DG units at various
locations. Moreover, the problem with the proposed character-
istic is formulated as a constrained NLP problem to determine
the optimal relay settings for each system. A comparative study
is presented that highlights the advantages of the proposed non-
standard characteristic over the conventional timecurrent relay
characteristics.
II. PROPOSED RELAY CHARACTERISTIC
Protection schemes based on DOCRs should provide protec-
tion against interphase faults as well as single-line-to-ground
(SLG) faults. Thus, in this paper, it is assumed that all of the sys-
tems under study are equipped with three-phase DOCRs that are
capable of providing protection against SLG, line-to-line (LL),
double-line-to-ground (DLG), and three phase faults [21]. Pro-
tection schemes involving both phase and ground relays (based
on zero- and negative-sequence currents), have not been con-
sidered in this paper.
Most of the conventional DOCRs used up till today utilize
a timecurrent characteristic where the time changes inversely
with current magnitude. The characteristic equation varies
slightly depending on the type of DOCR used and the standard,
which the relay manufacturer follows. In this study, the IEC
255-3 standard inverse characteristic is followed [22]. Since
microprocessor-based DOCRs provide an opportunity to de-
velop user-dened characteristics for the DOCRs overcurrent
unit [23], [24], this paper proposes a new nonstandard charac-
teristic for the DOCRs.
The proposed DOCR algorithm involves directional sensing
and tripping time computation. Usually, DOCR determines
the direction of the fault using the phase relationship of the
voltage and current. DOCRs are equipped with three voltage
transformers (VTs) for three-phase operation in addition to the
three existing current transformers (CTs) to measure voltage
and current, espectively [25]. Based on the measured quantities,
the respective voltages and currents magnitudes and phases are
computed. The algorithm used to determine the fault direction
varies among different relay manufacturers in terms of how
the angular relationships are sensed [25]. The 90 quadrature
method is typically used for direction detection. This method
involves comparing each currents phase angle with a polar-
izing quantity which is the line-to-line voltage between the
other two phases shifted by an angle commonly referred to as
the relay characteristic angle (RCA) for digital and numerical
relays [25][29].
Fig. 1. Proposed relay timecurrentvoltage characteristic with 2 and
0.1.
Since DOCRs have VTs already installed, this paper proposes
a modied DOCR characteristic for the tripping time compu-
tation that utilizes, in addition to the current, the fault voltage
magnitude at the DOCR. The proposed timecurrent-voltage
DOCR characteristic can be expressed as follows:
(1)
where is the tripping time (in seconds) of relay due to a
fault type occurrence at location . is the per-unit equiv-
alent of the phase fault voltage magnitude measured at relay
for a fault type at location , while is a constant parameter.
It is worth noting that the fault voltage magnitude used in the
proposed characteristic should not be used for direction com-
parisons. is the multiple of pickup currents which is equal
to . is the short-circuit current measured at the
secondary winding of the current transformer of relay for a
fault type at location while the pickup current is the min-
imum value of current above which the relay starts to operate.
is the time dial setting of relay . The constants and
are set to 0.14 and 0.02, respectively. Fig. 1 illustrates the char-
acteristic curve for a relay with and set to 2 and 0.1,
respectively.
The proposed characteristic integrates the fault voltage mag-
nitude as a third dimension to the conventional timecur-
rent characteristic which allows for additional possible conven-
tional tripping characteristics with only three settings to set and
optimize. The proposed DOCR characteristic consists of two
cascaded parts. The rst part is the conventional standard in-
verse DOCRoperating characteristic equation. This part ensures
maintaining the inverse timecurrent characteristic properties
with the common relay settings ( and ). The second part
of the proposed characteristic is the fault voltage magnitude ex-
ponential function with being the third setting. An exponen-
tial function has been chosen such that the following conditions
are fullled:
a) The proposed characteristic relies on the measure-
ment. varies depending on the fault location with re-
spect to the DOCRs location at which it is measured. This
measurement can reach zero for a bolted fault occurring
at the DOCRs VT terminal. Since the typical operation
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
SALEH et al.: OPTIMAL COORDINATION OF DIRECTIONAL OVERCURRENT RELAYS 3
Fig. 2. Effect of the setting on the proposed relay tripping time with
1.5 and 0.1.
time of the DOCR is about 1620 ms [30], the exponen-
tial function is used to ensure that the operation time does
not become zero when the fault voltage magnitude is zero.
b) The DOCR operating time is designed to be proportional
to the drop in voltage due to the fault. A large drop in
voltage signies that the fault is closer to the relay and,
thus, fast operation is needed and vice-versa. Thus, the
reciprocal of the exponential is considered with the expo-
nential superscript being .
c) The effect of is cancelled by setting to zero and
the conventional timecurrent characteristic is obtained
which makes the proposed approach quite exible.
d) The exponential part has been chosen such that it does not
alter the effect of and on the relay characteristic.
The rate of change in tripping time with respect to change in
is controlled by varying (Fig. 2). Thus, in the proposed
method, is considered as an additional setting for the relay
and is added to the coordination problem as an optimization
variable with the intention of attaining optimal settings for each
DOCR that solves the coordination problem. The parameter
can be either xed (the same for all relays in a given system)
or could be designed to have a different value for each DOCR.
Both cases are considered while formulating and solving the
coordination problem.
For a three-phase DOCR, three tripping times are computed
that correspond to each phase, respectively. This means that
the tripping time corresponding to phase will be calculated
based on phase voltage and current magnitudes ( for the
element). In a similar manner, the tripping times for phases
and are computed using the corresponding phase currents
and voltages (i.e., for the element, and for the
element).
In traditional relays, the element with the greatest current
magnitude operates since the relay characteristic is dependent
on the fault current magnitude only. This will not necessarily
be the same for the proposed characteristic since it depends on
voltage and current. The voltage magnitude setting, and cur-
rent magnitude will all simultaneously decide on the fastest op-
erating element. For any fault type, the algorithm will measure
the magnitude of each phase voltage and current. The operating
time for each pair of phase quantities ( and , and ,
and ) is calculated using (1), and the fastest element will
determine the overall operating time of the relay. This applies
to primary and backup relays. Thus, the results reported in this
paper present the fastest operating element.
III. PROTECTION COORDINATION PROBLEM FORMULATION
The protection coordination problem is developed for each
system in order to nd the optimal settings of each DOCR
for three cases considered in this study. Case 1 is basically
formulating and solving the problem using DOCRs with a
conventional standard inverse timecurrent characteristic,
while cases 2 and 3 use DOCRs equipped with the proposed
timecurrentvoltage characteristic. Case 2 is solved consid-
ering that all DOCRs will have the same value for while in
case 3, the optimization is designed such that each DOCR can
have a different value for the setting. In each case, all relays
were assumed to be identical where they are equipped with the
same tripping characteristic.
The main objective of the protection coordination optimiza-
tion (PCO) model is to minimize the sum of DOCRs operating
time ( ) (including the operating times of DOCRs due to SLG,
LL, DLG, and three phase faults occurring at near end, mid, and
far end on each line of a system) subject to protection coordina-
tion, relay setting, and relay operating time constraints. There-
fore, the objective function can be represented as follows:
Minimize (2)
where is the total number of relays with being the relay iden-
tier, is the total number of fault types studied with being
the fault type identier, and is the total number of fault loca-
tions investigated with being the fault-location identier. The
superscript represents primary relays, while represents the
backup relay with being the total number of backup relays
for each primary relay. The backup relay is dependent on the
primary relay for a fault type that occurred at location . The
variables and refer to the primary and backup relay
operating time for a fault of type at location , respectively.
The main optimization decision variables are the relay set-
tings which depend on the case considered. The pickup current
settings are predetermined based on the values given in [31]
and [32]. Thus, these values are considered as given parameters
in the PCO model. On the other hand, for DOCRs are
the variables of the PCO model in case 1. For case 2, the vari-
ables are and one additional variable . As for case 3,
the variables are , and for DOCRs. Therefore, case
1 is an LP problem since the conventional characteristic is lin-
earized (xed settings). However, cases 2 and 3 are an NLP
problem due to the nonlinearity of the proposed DOCR tripping
characteristic. Yet, both types of formulations have the same ob-
jective function and set of constraints.
The PCO model involves satisfying the protection coordina-
tion constraints. Each primary DOCR in a system requires a
backup DOCR to operate in the event of failing to isolate the
fault in its zone. Thus, the coordination time interval (CTI) is
the minimum gap in time between the operation of the primary
and backup DOCRs that must be maintained. In this paper, CTI
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
4 IEEE TRANSACTIONS ON POWER DELIVERY
Fig. 3. Power distribution system of the IEEE 14-bus system under study.
is considered to be 0.3 s. This constraint can be expressed as
follows:
(3)
Additional constraints are imposed on TDS settings and relay
operating time. These constraints can be expressed as
(4)
(5)
The lower and upper limits on the DOCRs TDS (
and ) are set to 0.05 and 1, respectively. Finally, the
operating time of each primary and backup DOCR for all fault
types and locations is constrained by a lower limit set to 20 ms
[30].
IV. SYSTEM DETAILS AND SIMULATION SETUP
This section describes the test systems under study. The main
study consists of a total of 16 possible system scenarios that are
generated based on IEEE 14-bus and IEEE 30-bus standard dis-
tribution networks by varying the DG location. The single-line
diagram of the IEEE 14-bus test system is shown in Fig. 3 [33].
This meshed distribution network is fed through two 60-MVA
132-kV/33-kVutility transformers with 10%transient reactance
connected at buses 1 and 2. The system is equipped with 16
DOCRs. Fig. 3 shows nodes (F8F15) where each one repre-
sents a set of faults on each line. For instance, F8 consists of a
set of bolted SLG, LL, DLG, and three-phase faults occurring at
near-end, midpoint, and far-end locations that are investigated
on the line between buses 1 and 4.
Based on this system, six scenarios were created where the
rst is the base scenario with no DG units installed, and the re-
maining ve scenarios are developed by adding DG units to the
system each time. As shown in Table I, the DG units utilized in
this study are each rated at 2MVA, synchronous type with 9.67%
transient reactance, and are connected to the system through a
480-V/33-kV stepup transformer with 5% transient reactance.
Each fault is associated with two primary DOCRs, one from
each side that are each backed up by other DOCRs. For example,
if a bolted three phase near-end (close to R1) fault occurs at node
, R1 and R2 are the two primary DOCRs. R1 is backed
up by R4 and R6 while R2 is backed up by R11. Thus, for each
TABLE I
DESCRIPTION OF SYSTEM SCENARIOS
Fig. 4. Power distribution system of the IEEE 30-bus system under study.
fault location, primary and backup DOCRs are dened and ac-
cordingly implemented in the PCO models.
Similarly, an additional nine system scenarios are generated
based on the IEEE 30-bus distribution network [33]. The stan-
dard single-line diagram of the system is shown in Fig. 4. The
base selected for all the systems under study is 100 MVA and 33
kV. A summary of the system scenarios investigated is given in
Table I.
The optimization model developed for each system scenario
for cases 2 and 3 is solved by using the MATLAB min-
imum-constrained nonlinear multivariable function [34]. This
built-in MATLAB function relies on the reduced gradient ap-
proach (rst-order optimality) for solving constrained nonlinear
optimization problems. There are several algorithms available
for solving a constrained nonlinear programming problem.
The sequential quadratic programming (SQP) algorithm is
chosen to solve the PCO models [18], [35]. A full description
of the MATLAB SQP algorithm is provided in [34]. On the
other hand, the PCO model that has been developed for each
scenario, for case 1, is solved using the Simplex technique.
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
SALEH et al.: OPTIMAL COORDINATION OF DIRECTIONAL OVERCURRENT RELAYS 5
TABLE II
OPTIMAL , and CONSIDERING CASES 13
V. OPTIMAL RELAY SETTINGS USING THE PROPOSED
TIMECURRENTVOLTAGE CHARACTERISTIC
The protection coordination problem presented in Section III
is solved by considering the conventional DOCR timecurrent
characteristic and the proposed DOCR timecurrentvoltage
characteristic. The optimal DOCR settings as well as operating
times are determined for all of the presented 16 system sce-
narios. Comprehensive analysis of the results obtained for all
of the investigated system scenarios is presented.
A. Simulation Results of the IEEE 14-bus System With 5 DGs
In this subsection, the simulation results obtained for system
scenario 6 (refer to Table I), which consists of ve synchronous-
based DG units connected at buses 3 7 on the IEEE 14-bus
power distribution network, are analyzed. The optimal values of
and as well as the objective function values consid-
ering the conventional characteristic (case 1) and the proposed
relay characteristic (cases 2 and 3) are presented in Table II.
As can be seen from the results, the relay setting is xed
for all cases while the TDS values have changed when consid-
ering the proposed characteristic. The optimal DOCR settings
achieved in each case ensure proper coordination of the DOCRs
under SLG, DLG, LL, or three-phase faults that can occur at the
near-end, mid, or far-end location on each line of the system.
Moreover, the total relay operating time is reduced by 28.4%
in case 2 (considering the same setting for all relays) with
respect to case 1 (conventional characteristic). Further reduc-
tion in (38.7% compared to case 1) is achieved in case 3
where the optimal value of is determined for each DOCR.
As can be seen, the new proposed timecurrentvoltage char-
acteristic is capable of achieving better overall relay operating
time compared to the conventional timecurrent characteristic.
For brevity, a breakdown of the DOCR operation times due to
three-phase, LL, DLG, and SLG faults at the midpoint on each
line are given in Tables IIIVI considering the conventional case
and the case where can vary for each relay (case 3), respec-
tively. The operating time of each primary and backup DOCR
TABLE III
OPTIMAL PRIMARY AND BACKUP RELAY OPERATING TIMES OF SCENARIO 6
- CASE 1/CASE 3 DUE TO THREE-PHASE FAULTS
TABLE IV
OPTIMAL PRIMARY AND BACKUP RELAY OPERATING TIMES OF SCENARIO 6 -
CASE 1/CASE 3 DUE TO LL FAULTS
satises the relay operating time constraint. It is also observed
that all primary-backup relays are correctly coordinated with a
minimum value of 300 ms for CTI. Also, it can be seen that
for all fault types, the relay tripping times obtained from the
proposed relay characteristic are always faster than the times
attained from the conventional relay characteristic. Similar ob-
servations have been noticed for other scenarios.
B. Comprehensive Analysis
In order to test the performance of the proposed DOCR
characteristic, the problem formulation presented in Section III
is applied to the 16 scenarios given in Table I with each relay
equipped with the proposed timecurrentvoltage character-
istic. The overall minimum operating time for the three
cases that are implemented on the 16 system scenarios is
summarized in Table VII. In general, it can be seen that the
proposed timecurrentvoltage characteristic (both cases 2
and 3) achieves a better overall minimum operating time in
all scenarios. For the cases where no DG units are present in
the system (system scenario 1 for the IEEE 14-bus system
and 7 for the IEEE 30-bus system), in comparison to case 1, a
reduction in of 23.2% and 11% was achieved in case 2 while
33% and 21.5% were achieved in case 3, respectively. Thus,
the proposed relay characteristic outperforms the conventional
characteristic in cases where no DG is present. In addition,
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
6 IEEE TRANSACTIONS ON POWER DELIVERY
TABLE V
OPTIMAL PRIMARY AND BACKUP RELAY OPERATING TIMES OF SCENARIO 6 -
CASE 1/CASE 3 DUE TO DLG FAULTS
TABLE VI
OPTIMAL PRIMARY AND BACKUP RELAY OPERATING TIMES OF SCENARIO 6 -
CASE 1/CASE 3 DUE TO SLG FAULTS
irrespective of the number of DG units, in the system size
as well as DG location, the proposed relay characteristic can
achieve signicantly better DOCR operating time. For the
IEEE 14-bus system, the percentage reduction in varies from
21.3% to 33.6% in case 2 whereas it is 28.7% to 48% in case
3. On the other hand, percentage reductions in range from
4.69% to 11.4% in case 2 while 20.7% to 27.6% in case 3 were
obtained for the IEEE 30-bus system. The amount of reduction
in is dependent on the number of DGs as well as location.
Table VIII shows the optimal value obtained in case 2
(considering the same for all relays) when solving the PCO
model for each system scenario. The optimal value for the
IEEE 14-bus system (system 1 up to 6) roughly varies around
the value of 1 while for the IEEE 30-bus system, it is limited to
a small value around zero. As can be seen, the setting varies
depending on the systemscenario and can be optimally set using
the proposed model. Small values limit the fault voltage mag-
nitude effect and, thus, limit the amount of reduction achieved
with respect to case 1. Allowing each relay to have its own op-
timal can achieve higher reduction in . It is worth noting
that setting a xed value of for all relays versus allowing the
value of to be different for each relay will depend on the pro-
tection engineer designer preference.
TABLE VII
TOTAL RELAYS OPERATING TIME FOR ALL CASES AND SCENARIOS
TABLE VIII
OPTIMAL FOR CASE 2ALL SCENARIOS
TABLE IX
PERCENTAGE REDUCTION IN TIME WITH RESPECT
TO CASE 1 FOR DIFFERENT SYSTEM SCENARIOS
C. Performance of the Proposed Characteristic in the
Presence of the Inverter-Based DG
The impact of inverter-based and synchronous-based DG
units on the amount of percentage reduction that can be
achieved using the proposed characteristic is investigated in
this subsection. Four system scenarios were implemented con-
sidering to be xed for all relays (case 2) and variable (case
3). Each system scenario has either inverter-based DG units or a
combination of inverter-based and synchronous-based DGunits
connected to the IEEE 30-bus network. The synchronous-based
DG units are rated at 2 MVA with the specications mentioned
previously. On the other hand, the inverter-based DG units are
rated at 2 MVA and are controlled in the constant current mode.
Thus, they are modeled as ideal current sources injecting 2
p.u. of the DG-rated capacity during fault conditions. Table IX
shows the four system scenarios and the corresponding re-
sulting percentage reduction in total relays operating time
with respect to case 1.
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
SALEH et al.: OPTIMAL COORDINATION OF DIRECTIONAL OVERCURRENT RELAYS 7
TABLE X
PRECENTAGE REDUCTION IN TIME WITH RESPECT TO CASE 1
FOR DIFFERENT DG CAPACITIES AND LOCATIONS
The resulting percentage reduction in conrms that re-
gardless of type of DG unit connected to a network, the pro-
posed relay characteristic (case 2 and 3) will achieve signi-
cant improvement with respect to the conventional character-
istic. Moreover, the amount of percentage reduction is affected
by the type of DG installed.
D. Impact of DG Capacity and Location
Since the synchronous-based DG unit has a more profound
impact on short-circuit current, this subsection analyzes the im-
pact of the synchronous-based DG units capacity and location
on the amount of percentage reduction that can be achieved
using the proposed relay characteristic. Multiple simulations
were performed considering to be xed for all relays (case2)
and variable (case3). For each simulation, the DG is installed
at a specic location while varying its capacity. For brevity, for
each system, three locations are investigated for DG capacities
of 2 MVA, 5 MVA, and 9 MVA. Table X demonstrates the 18
simulation setups and the corresponding resulting percentage
reduction in total relays operating time with respect to case 1.
From the results, three main observations can be deduced.
First, case 3 achieves a higher percentage reduction in with
respect to case 1 than case 2 in all simulation setups. Second,
it is clear that the amount of percentage reduction depends on
the network (IEEE 14 bus or IEEE 30 bus) to which the new
proposed DOCRs are connected. Finally, it is observed that de-
pending on the location and capacity of DG in the network, the
amount of percentage reduction experiences slight variation in
cases 2 and 3. Although there is no general trend to the amount
of reduction with respect to an increase in DGcapacity at a given
location, the proposed relay characteristic always outperforms
the conventional timecurrent characteristic and, thus, can re-
duce the overall relays operating time for all cases.
It is worth noting that the analysis performed in this paper is
based on a xed network topology. Power systems may expe-
rience changes in the network topology due to various reasons,
such as network recongurations, maintenance activities, and
isolation of faulted sections [8]. In such occasions, some of the
primary/backup DOCR pairs for specic faults may not be co-
ordinated correctly. According to [8], [9], [13], and [31], proper
coordination can be achieved by modifying the PCO model to
include additional sets of coordination constraints that consider
all network topologies due to single-line outage contingencies
in a system. This will result in optimal DOCR settings that are
different from the optimal DOCR settings achieved under the
assumption of xed network topology. In addition to the above
concern, the fault current direction might change with changes
in the network topology. This issue was not addressed in pre-
vious work focusing on protection coordination. In this paper,
the effect of dynamic changes in the network topology has not
been considered.
VI. CONCLUSION
This paper proposes a new timecurrentvoltage relay char-
acteristic that can be utilized by DOCRs in a meshed distribu-
tion network in the presence of DG units. The proposed char-
acteristic relies on utilizing the fault voltage magnitude in addi-
tion to the current to determine the operating time of DOCRs.
In addition to the commonly used relays settings, a third relay
setting , responsible for tuning the effect of voltage on relay
operating time, was introduced. This setting can either have a
xed common value for all DOCRs or each DOCR can have
its own optimal setting value. The protection coordination
problem is formulated and solved considering the conventional
and proposed relay characteristic. The optimal settings attained
by solving the PCO model for cases with either a conventional
or proposed DOCR characteristic ensure proper coordination of
all DCORs under SLG, DLG, LL, and three-phase faults that
might occur at near-end, mid, and far-end locations on each line
of a system. The results showthat the utilization of the proposed
timecurrentvoltage characteristic for each DOCRin a meshed
distribution network can achieve a signicant reduction in the
total relay operating time in the absence and presence of syn-
chronous-based and inverter-based DG. The amount of reduc-
tion will depend on the system structure, type, number, and ca-
pacity as well as the location of DG units. In addition, by setting
each relay with its optimal settings (as opposed to having the
same value of for all relays), further reduction in the overall
relays operating time can be achieved.
REFERENCES
[1] P. Sutherland, Protective device coordination in an industrial power
system with multiple sources, IEEE Trans. Ind. Appl., vol. 33, no. 4,
pp. 10961103, Jul. 1997.
[2] J.-H. Teng, Unsymmetrical short-circuit fault analysis for weakly
meshed distribution systems, IEEE Trans. Power Syst., vol. 25, no.
1, pp. 96105, Feb. 2010.
[3] B. Hussain, S. Sharkh, and S. Hussain, Impact studies of distributed
generation on power quality and protection setup of an existing dis-
tribution network, in Proc. IEEE Int. Symp. Power Electron. Elect.
Drives Autom. Motion, 2010, pp. 12431246.
[4] W. El-khattam and T. Sidhu, Restoration of directional overcurrent
relay coordination in distributed generation systems utilizing fault cur-
rent limiter, IEEE Trans. Power Del., vol. 23, no. 2, pp. 576585, Apr.
2008.
[5] W. Najy, H. Zeineldin, and W. Woon, Optimal protection coordina-
tion for microgrids with grid-connected and islanded capability, IEEE
Trans. Ind. Electron., vol. 60, no. 4, pp. 16681677, Apr. 2013.
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
8 IEEE TRANSACTIONS ON POWER DELIVERY
[6] V. Pandi, H. Zeineldin, and W. Xiao, Determining optimal location
and size of distributed generation resources considering harmonic and
protection coordination limits, IEEE Trans. Power Syst., vol. 28, no.
2, pp. 12451254, May 2013.
[7] H. Zeineldin, Y.-R. Mohamed, V. Khadkikar, and V. Pandi, A protec-
tion coordination index for evaluating distributed generation impacts
on protection for meshed distribution systems, IEEE Trans. Smart
Grid, vol. 4, no. 3, pp. 15231532, Sep. 2013.
[8] A. S. Noghabi, J. Sadeh, and H. R. Mashhadi, Considering different
network topologies in optimal overcurrent relay coordination using a
hybrid ga, IEEE Trans. Power Del., vol. 24, no. 4, pp. 18571863,
Oct. 2009.
[9] A. Urdaneta, R. Nadira, and L. Perez Jimenez, Optimal coordination
of directional overcurrent relays in interconnected power systems,
IEEE Trans. Power Del., vol. 3, no. 3, pp. 903911, Jul. 1988.
[10] A. Urdaneta, H. Restrepo, S. Marquez, and J. Sanchez, Coordina-
tion of directional overcurrent relay timing using linear programming,
IEEE Trans. Power Del., vol. 11, no. 1, pp. 122129, Jan. 1996.
[11] A. Urdaneta, L. Perez, J. Gomez, B. Feijoo, and M. Gonzalez, Pre-
solve analysis and interior point solutions of the linear programming
coordination problem of directional overcurrent relays, Int. J. Elect.
Power Energy Syst., vol. 23, no. 8, pp. 819825, 2001.
[12] B. Chattopadhyay, M. Sachdev, and T. Sidhu, An on-line relay coor-
dination algorithm for adaptive protection using linear programming
technique, IEEE Trans. Power Del., vol. 11, no. 1, pp. 165173, Jan.
1996.
[13] A. Urdaneta, L. Perez, and H. Restrepo, Optimal coordination of
directional overcurrent relays considering dynamic changes in the
network topology, IEEE Trans. Power Del., vol. 12, no. 4, pp.
14581464, Oct. 1997.
[14] C. W. So and K. K. Li, Time coordination method for power system
protection by evolutionary algorithm, IEEE Trans. Ind. Appl., vol. 36,
no. 5, pp. 12351240, Sep. 2000.
[15] P. Bedekar, S. Bhide, and V. Kale, Optimum coordination of overcur-
rent relays in distribution system using genetic algorithm, in Proc. Int.
Conf. Power Syst., Dec. 2009, pp. 16.
[16] M. M. Mansour, S. F. Mekhamer, and N. EL-SHERIF EL-KHAR-
BAWE, A modied particle swarm optimizer for the coordination of
directional overcurrent relays, IEEE Trans. Power Del., vol. 22, no.
3, pp. 14001410, Jul. 2007.
[17] H. Zeineldin, E. El-Saadany, and M. Salama, Optimal coordination
of overcurrent relays using a modied particle swarm optimization,
Elect. Power Syst. Res., vol. 76, no. 11, pp. 988995, 2006.
[18] D. Birla, R. Maheshwari, and H. Gupta, A new nonlinear directional
overcurrent relay coordination technique, banes and boons of near-end
faults based approach, IEEE Trans. Power Del., vol. 21, no. 3, pp.
11761182, Jul. 2006.
[19] T. Keil and J. Jager, Advanced coordination method for overcurrent
protection relays using nonstandard tripping characteristics, IEEE
Trans. Power Del., vol. 23, no. 1, pp. 5257, Jan. 2008.
[20] N. Nimpitiwan, G. Heydt, R. Ayyanar, and S. Suryanarayanan, Fault
current contribution from synchronous machine and inverter based
distributed generators, IEEE Trans. Power Del., vol. 22, no. 1, pp.
634641, Jan. 2007.
[21] C. R. Mason, The Art and Science of Protective Relaying. Hoboken,
NJ, USA: Wiley, 2012.
[22] Single Input Energizing Quality Measuring Relays With Dependent or
Independent, IEC Publ. 255-3 (1989-05), 1989.
[23] T. Areva, Network Protection & Automation Guide. Barcelona,
Spain: Cayfosa, 2002, pp. 9295, layout by Flash Espace, Montpellier,
France.
[24] Siprotec Multi-function Protective Relay With Local Control
7sj62/63/64 v4. 6 Manual, ver. Order No. C53000-G1140-C147-7,
Siemens AG.
[25] J. Horak, Directional overcurrent relaying (67) concepts, in Proc.
IEEE 59th Annu. Protect. Relay Eng. Conf., 2006, p. 13.
[26] A. Bakar, H. Mokhlis, H. Illias, and P. Chong, The study of directional
overcurrent relay and directional earth-fault protection application for
33 kv underground cable system in Malaysia, Int. J. Elect. Power En-
ergy Syst., vol. 40, no. 1, pp. 113119, 2012.
[27] K. Zimmerman and D. Costello, Fundamentals and improvements
for directional relays, in Proc. IEEE 63rd Annu. Conf. Protect. Relay
Eng., 2010, pp. 112.
[28] Instruction Manual, Directional Overcurrent Protection
RelayGRD140, Toshiba, 2004.
[29] Directional/non Directional Overcurrent Protection. Technical Data
Sheet (P12xy/EN TDS/H76), AREVA, 2012.
[30] J. L. Blackburn and T. J. Domin, Protective Relaying: Principles and
Applications, 3rd ed. Boca Raton, FL, USA: CRC, 2007.
[31] A. S. Noghabi, H. R. Mashhadi, and J. Sadeh, Optimal coordination of
directional overcurrent relays considering different network topologies
using interval linear programming, IEEE Trans. Power Del., vol. 25,
no. 3, pp. 13481354, Jul. 2010.
[32] M. Ezzeddine, R. Kaczmarek, and M. Iftikhar, Coordination of direc-
tional overcurrent relays using a novel method to select their settings,
IET Gen., Transm. Distrib., vol. 5, no. 7, pp. 743750, 2011.
[33] Univ. Washington, Seattle, WA, USA, Power systems test case
archive, Tech Rep., 2006. [Online]. Available: http://www.ee.wash-
ington.edu/research/pstca
[34] M. O. Toolbox, Optimization Toolbox. Natick, MA, USA: The Math-
works Inc., 2002.
[35] P. P. Bedekar and S. R. Bhide, Optimum coordination of directional
overcurrent relays using the hybrid ga-nlp approach, IEEE Trans.
Power Del., vol. 26, no. 1, pp. 109119, Jan. 2011.
Khaled A. Saleh (S12) was born in Cairo, Egypt, in 1991. He received the
B.Sc. degree in electrical engineering fromThe Petroleum Institute, Abu Dhabi,
United Arab Emirates (UAE), in 2012, the M.Sc. degree in electrical power en-
gineering from Masdar Institute of Science and Technology, Abu Dhabi, UAE,
in 2014, and is currently pursuing the Ph.D. degree in electrical and computer
engineering from University of Waterloo, Waterloo, ON, Canada.
His current research interests include power system protection, distributed
generation, system reconguration, and optimization techniques.
H. H. Zeineldin (M06SM13) received the B.Sc. and M.Sc. degrees in elec-
trical engineering from Cairo University, Cairo, Egypt, in 1999 and 2002, re-
spectively, and the Ph.D. degree in electrical and computer engineering from
the University of Waterloo, Waterloo, ON, Canada, in 2006.
He was with Smith and Andersen Electrical Engineering, Inc., where he was
involved with projects involving distribution system design, protection, and dis-
tributed generation. He was then a Visiting Professor at the Massachusetts In-
stitute of Technology, Cambridge, MA, USA. He is with the Faculty of Engi-
neering, Cairo University, and currently he is an Associate Professor with the
Masdar Institute of Science and Technology, Masdar City, UAE. He is currently
an Editor of the IEEE TRANSACTIONS ON ENERGY CONVERSION and IEEE
TRANSACTIONS ON SMART GRIDS. His current interests include power system
protection, distributed generation, and deregulation.
A. Al-Hinai (M08SM11) received the B.Sc. degree in electrical engineering
from Sultan Qaboos University, Muscat, Oman, in 1997. He received the M.Sc.
and Ph.D. degrees in electrical engineering from West Virginia University, Mor-
gantown, WV, USA, in 2000 and 2005, respectively.
He joined Sultan Qaboos University as a Demonstrator in 1998. He is an
Assistant Professor in the Electrical and Computer Engineering Department,
Sultan Qaboos University. Currently, he is a Visiting Professor at Masdar In-
stitute of Science and Technology, Abu Dhabi, United Arab Emirates. In addi-
tion, he is holding a member position at the Authority of Electricity Regulation,
Oman. His areas of interest are power systems, such as control and operation,
along with multiagent systems, applications for the control of power systems,
optimization techniques, and applications for power systems.
Ehab F. El-Saadany (SM05) was born in Cairo, Egypt, in 1964. He received
the B.Sc. and M.Sc. degrees in electrical engineering from Ain Shams Univer-
sity, Cairo, Egypt, in 1986 and 1990, respectively, and the Ph.D. degree in elec-
trical engineering from the University of Waterloo, Waterloo, ON, Canada, in
1998.
Currently, he is a Professor with the Department of Electrical and Computer
Engineering, University of Waterloo. His current research interests include dis-
tribution system operation and control, distributed generation, smart-grid ap-
plications, self-healing mechanisms, power quality, and microelectromechan-
ical-systems micropower generators.
Dr. El-Saadany is the holder of the Canada Research Chair in Smart Distri-
bution Systems in recognition of his contributions to the areas of distributed
generation and microgrids.