IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 28, NO.
2, MAY 2013
Determining Optimal Location and Size of
Distributed Generation Resources Considering
Harmonic and Protection Coordination Limits
V. Ravikumar Pandi, Member, IEEE, H. H. Zeineldin, Member, IEEE, and Weidong Xiao, Member, IEEE
Abstract—In this paper, a new optimization problem is proposed
to determine the maximum distributed generation (DG) pene-
tration level by optimally selecting types, locations and sizes of
utility owned DG units. The DG penetration level could be limited
by harmonic distortion because of the nonlinear current injected
by inverter-based DG units and also protection coordination
constraints because of the variation in fault current caused by
synchronous-based DG units. Hence the objective of the proposed
problem is to maximize DG penetration level from both types
of DG units, taking into account power balance constraints, bus
voltage limits, total and individual harmonic distortion limits
specified by the IEEE-519 standard, over-current relay operating
time limits, and protection coordination constraints. The DG
penetration study is formulated as a nonlinear programming
(NLP) problem and tested on the IEEE-30 bus looped distribution
network with ten load and DG scenarios. Similarly, feasibility
assessment of customer owned DG unit installations considering
power quality and protection coordination is also studied. Sim-
ulation results show the effectiveness of the proposed approach,
which can serve as an efficient planning tool for utility operators.
Index Terms—Distributed generation, harmonic distortion,
harmonic power flow, particle swarm optimization, protection
coordination.
I. INTRODUCTION
R ECENT advancement in research and development on
distributed generation (DG) technology such as solar,
wind, biomass, fuel cell and micro-turbines has created an
important role for DG in the future energy supply due to its
improved performance, reliability and flexibility to achieve
higher energy efficiency and reduced emission [1], [2]. The
DG units are interconnected with the utility grid as per the
IEEE-1547 standard [3], [4]. The location, type, and size of DG
units are the three main factors that can affect the fault current
levels, protection coordination of relays, stability, and power
quality [5]–[7] and this can limit the amount of DG penetration.
The DG units may be owned either by utility or by customers
[8], [9]. In the case of a utility owned DG installation, the utility
has to optimally plan the location and size of the DG units in
order to improve network benefits and reliability [10]–[12].
In practice, one is not always able to site DG at the locations
Manuscript received January 23, 2012; revised June 03, 2012; accepted July
15, 2012. Date of publication August 13, 2012; date of current version April 18,
2013. This work was supported and funded by the Masdar Institute of Science
and Technology. Paper no. TPWRS-00068-2012.
The authors are with the Masdar Institute of Science and Tech-
nology, Abu Dhabi, United Arab Emirates (e-mail: vpandi@masdar.ac.ae;
hzainaldin@masdar.ac.ae; mwxiao@masdar.ac.ae).
Digital Object Identifier 10.1109/TPWRS.2012.2209687
0885-8950/$31.00 © 2012 IEEE
1245
determined by an optimization algorithm, however it gives
the planner an idea of where DG might be the most beneficial
[13]. In the case of customer owned DG installation, the utility
planner should conduct a feasibility and assessment study to
evaluate any technical issues resulting from the new installation
of customer owned DG installation [14].
During the last decade, power electronic converters have un-
dergone a fast evolution due to the development of fast semi-
conductor switches and also the introduction of real-time con-
trollers that can implement advanced and complex control algo-
rithms efficiently [15]. DG units of the inverter-based type tend
to have more impact on the system harmonic levels than syn-
chronous-based DG. The limitations on current and voltage har-
monics, which can cause undesirable effects on various power
system equipments and the measurement and equipment mod-
elling for harmonic analysis have been given in the IEEE-519
standard [16]. The penetration level of photovoltaic generation
in radial distribution system considering the limits on voltage
magnitudes and conductor current flow is investigated in [17].
The maximum DG penetration level based on the individual har-
monic limit is examined for a simple system in [18]. This ap-
proach uses 7th and 9th order individual current harmonic limits
given by the IEEE-519 standard and proves that 100% DG pen-
etration is theoretically possible.
On the other hand, synchronous-based DG units has a much
more profound effect on protection coordination than inverter-
based DG [19]–[21]. In [22], [23], fault current limiters are im-
plemented to mitigate the impact of synchronous-based DG on
protection coordination. In [24], the maximum DG capacity for
radial distribution systems was calculated considering coordi-
nation between recloser and fuse to avoid reliability degrada-
tion. Optimal placement of synchronous-based DG, considering
the impact on short circuit index level, is solved using particle
swarm optimization [25]. An analytical approach to minimize
the system losses by installing various type of DG units (PV,
wind, fuel cell, and micro turbine based on real and reactive
power injection only) is discussed in [26]. However a compre-
hensive study considering DG type, size, and location simulta-
neously for maximizing the DG penetration has not been ad-
dressed in the literature.
In this paper, an optimization framework is proposed to max-
imize the DG penetration level in looped distribution systems
taking into account power balance, voltage limits, harmonic
limits and protection constraints. Since the harmonic distortion
is caused by inverter-based DG units and protection coordina-
tion failure of directional over-current relays (OCR) are due
to the synchronous-based DG units, an algorithm that includes
1246
Fig. 1. Framework showing fitness evaluation sequence.
both decoupled harmonic power flow (DHPF) for harmonic es-
timation and short circuit analysis for protection coordination
is embedded within the optimization framework. In the case
of customer owned DG installation, the feasibility assessment
studies are conducted by the utility planner with the support
of the aforementioned optimization tool. The problem is solved
using the particle swarm optimization (PSO) algorithm due to
the nonlinear and nonconvex nature. The problem is tested on
the standard IEEE-30 bus distribution system with ten load and
DG scenarios and the maximum penetration level is determined
considering both types of DG units and compared with the case
of optimizing DG penetration with only inverter-based DG.
II. PROBLEM FORMULATION
The proposed problem is formulated to maximize the overall
DG penetration level by optimally selecting the location and
size of both inverter-based and synchronous-based DG units
using particle swarm optimization (PSO). The constraints of the
proposed problem include fundamental frequency real and re-
active power balance, RMS voltage limits, individual and total
voltage harmonic limits at each bus, relay operating time limits
and protection coordination constraints. The fitness evaluation
sequence for each PSO candidate solution is shown in Fig. 1.
Initially, the Newton-Raphson based conventional power
flow is conducted at fundamental frequency to calculate the
fundamental voltage components considering both types of DG
units, taking into account the power balance constraints. After
running conventional power flow, the sequence has been di-
vided into two parallel subroutines namely decoupled harmonic
power flow for harmonic analysis and short circuit studies for
protection coordination analysis. In the first subroutine, the
result of conventional power flow is used by the DHPF algo-
rithm which estimates the higher order harmonic components
with inverter-based DG units. In the second subroutine, the
impact of DG units on fault current is considered to optimize
the relay settings with protection coordination constraints. The
details of the proposed problem formulation are described in
the following subsections.
A. Harmonic Power Flow Analysis
Harmonic power flow techniques are useful to estimate the
system distortion due to the presence of nonlinear devices [27],
[28]. Among various harmonic power flow formulations, the
DHPF method is most commonly used due to its simplicity
[29]. The subroutine steps involving the decoupled harmonic
power flow solution are shown in Fig. 2. Initially, the results
IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 28, NO. 2, MAY 2013
Fig. 2. Flowchart of the decoupled harmonic power flow (Subroutine1).
of conventional power flow, system data and DG type, loca-
tion and capacity are inputted to DHPF subroutine. In higher
order harmonic frequencies, the transmission lines, shunt ca-
pacitors, synchronous machines, and linear loads are modeled
as an equivalent admittances using the results of the conven-
tional power flow and then a new admittance matrix is formu-
lated. Inverter-based DG units are modeled as harmonic current
injecting sources in the DHPF method. On the other hand, the
synchronous-based DG units are modeled as a source behind
an impedance. Nodal equations are solved for each individual
harmonic order to obtain the harmonic voltage and the result
is used to calculate the harmonic distortion [29], [30]. The har-
monic admittances of various components can be expressed as
follows:
(1)
(2)
(3)
where
fundamental real and reactive power demand
at bus ;
magnitude of fundamental voltage at bus ;
admittance of load connected at bus for th
order harmonic;
fundamental frequency admittance of the
capacitor connected at bus ;
shunt capacitor admittance at th order;
admittance of branch connected between buses
and at th order.
The harmonic admittances of various elements are used to
formulate harmonic admittance matrix . The fundamental
PANDI et al.: DETERMINING OPTIMAL LOCATION AND SIZE OF DISTRIBUTED GENERATION RESOURCES 1247

current injected by inverter-based DG units at bus and

the th harmonic order current are given by

(4)

(5)

where
fundamental real and reactive power
generation by inverter-based DG units
connected at bus ;
ratio of th harmonic order current to its Fig. 3. Flowchart of the protection relay coordination (Subroutine 2).
fundamental value.
Finally, a set of nodal equations are solved for estimating the
harmonic voltage profile values using (6) impedance matrix is formulated and three phase short circuit
currents are estimated in presence of DG units. The above
(6) constrained nonlinear optimization problem is solved using
gradient based approaches available in Matlab optimization
toolbox [31]. The optimal settings of TDS and are used for
B. Protection Coordination
calculating the penalty in constraint violations (refer to Fig. 1).
In looped distribution system, the directional over-current
relay (OCR) determines the direction of fault by relating phase C. Objective Function
angle of voltage and current vectors. Since inverter-based DG The objective is to maximize the DG penetration level with
units have less impact on short circuit current levels (short cir- respect to total system capacity. Since DG units deliver only
cuit current levels are approximately 1 p.u.), the short circuit real power as recommended by the IEEE-1547 standard [3], the
contribution of inverter-based DG unit is neglected by model- objective function can be defined as follows:
ling the inverter-based DG as an open circuit in the short circuit
analysis. On the other hand, synchronous-based DG units are
modeled using the subtransient reactance. An inverse time over % (9)
current characteristics is considered for the OCR relays as given
in the following: where

(7) total number of buses in the system;

real power generation from inverter and
synchronous based DG units;
where A and B are constants that depend on the type of OCR
characteristics. is the pickup current of the relay above total system MVA.
which it starts operating and is the short circuit fault cur-
rent flowing through the relay because of fault at location . D. Equality Constraints
is the time dial setting of OCR. The objective is to op- The real and reactive power balance constraints at funda-
timally determine the TDS and of each OCR unit in order mental frequency for each system bus can be given as follows:
to minimize both primary and back-up relay operating times as
follows:

(8) (10)

where is the relay index, is the total number of relays, is

the fault location index and is the total number of fault loca-
tion considered in the analysis. The primary and backup relay (11)
operating times are represented by and , respectively.
The optimal relay settings are obtained by minimizing the
where
total operating time as given in (8) taking into consideration
limits on operating time, time dial setting and pick up current, fundamental real and reactive power generation
which will be discussed in Section II-E4. The solution steps at bus ;
involved in the protection coordination problem is shown in
magnitude of th element of the fundamental
Fig. 3. The conventional power flow results and DG location
bus admittance matrix;
and capacity are considered as input to this subroutine. The bus
1248
angle of th element of the fundamental bus
admittance matrix;
fundamental voltage angle at bus .
E. Inequality Constraints
1) Bus Voltage Limits: The rms voltage magnitude at each
bus is bound by a specified lower and upper limit as follows
[32]:
(12)
where are the lower and upper rms voltage limits
which are set at 0.9 p.u. and 1.1 p.u., respectively.
2) Total Harmonic Distortion Limits: The IEEE-519 stan-
dard imposes limits on the total voltage harmonic distortion at
each bus as follows:
% (13)
where is the maximum permissible total voltage har-
monic distortion which is usually set at 5% [16].
3) Individual Harmonic Distortion Limits: The IEEE-519
standard also imposes limits on individual voltage harmonic dis-
tortion at each bus which can be expressed as follows:
% (14)
where is the maximum allowable voltage harmonic
distortion level at harmonic order and is considered to be equal
to 3% [16].
4) Protection Coordination Limits: The penetration of DG
units in the distribution system could affect the short circuit cur-
rent. It can be noticed from (7) that the relay operating time
is a function of the short circuit current. The magnitude of the
short circuit current will depend on the DG location and capacity
which can be represented as follows:
(15)
The coordination time interval (CTI) is the minimum time be-
tween primary and back up relay operating time which must be
maintained to guarantee the protection coordination. The limits
on relay operating time and coordination time is given by
(16)
(17)
where is the maximum relay operating time of all relays.
The relay TDS and pickup current settings are limited by min-
imum and maximum limits as given in the following:
(18)
IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 28, NO. 2, MAY 2013
(19)
where is the minimum limits on pick up current set point
chosen using the knowledge of steady state current flowing
through the th relay estimated using conventional load flow
and is the maximum limit on pick up current set point
chosen based on short circuit currents. and
are the minimum and maximum limits on time dial set points
set to 0.1 and 3.0, respectively.
III. APPLICATION OF PSO TO MAXIMIZE DG PENETRATION
Particle swarm optimization (PSO) is a population based
stochastic search technique, well known among evolutionary
paradigms, which optimizes a problem iteratively to improve
a candidate solution. It is inspired by social behavior of bird
flocking and fish schooling [33]. Similar to genetic algorithms,
it starts with a random population matrix in the search space;
however, PSO has no crossover and mutation operators. Each
candidate solution, called particle, in the population travels
with a velocity and it is bounded by its maximum velocity. At
each time step, velocity of each particle is updated by its own
flying experience and the best particle’s position in the
swarm and then particle position is also updated [34]:
(20)
(21)
where
velocity of th particle’s dimension at iteration
;
position of th particle’s dimension at iteration
;
inertia weight factor;
cognitive and social coefficients equals to 2.1;
random numbers in the range of [0,1];
best position of th particle at th iteration;
global best position of the swarm at th iteration.
In order to have better balance between global exploration
and local exploitation, the inertia weight is decreased linearly
with the number of iterations [35]. The inertia weight is updated
at each iteration as follows:
(22)
where
minimum and maximum values of inertia
weight set to be 0.4 and 0.9, respectively;
maximum iteration number.
The implementation of PSO algorithm to the optimal DG pen-
etration problem starts with the proper encoding of optimiza-
tion parameter. The real power generation of inverter-based and
PANDI et al.: DETERMINING OPTIMAL LOCATION AND SIZE OF DISTRIBUTED GENERATION RESOURCES 1249

synchronous-based DG units connected to the set of candidate

buses are considered as the parameter to be optimized,
hence the particle structure is defined as

(23)

Each candidate solution from PSO is passed to the conven-

tional power flow followed by the DHPF and protection coor-
dination subroutines. The various constraints on RMS voltage,
harmonic distortion, OCR operating time and coordination time
are verified for feasibility and if violation is found, those con-
straints are added to the objective using the penalty function
method. The application of PSO algorithm for finding the op-
timal DG penetration level with harmonics and protection con-
straints is shown in Fig. 4. The overall step-by-step procedure
can also be given as follows.
Step 1) Input system bus and branch data, and DG candidate
locations.
Step 2) Randomly initialize population of individuals and
the velocity vector.
Step 3) Run conventional power flow for each individual
to solve real and reactive power balance equations
taking into account both types of DG sources at fun-
damental frequency.
Step 4) Calculate the harmonic admittance matrix for each
harmonic order . Calculate the current injection
matrix taking into account the inverter-based DG
harmonic spectrum.
Step 5) Run DHPF algorithm to estimate the harmonic
voltage components using (6).
Step 6) Run protection coordination subroutine to estimate
the optimal pick-up current and time dial settings as
explained in Section II-B.
Step 7) Evaluate objective function of each individual par-
ticle and add penalty function if there is any con-
straint violation.
Step 8) Find and vectors. Fig. 4. Flowchart of PSO-based optimal DG penetration.
Step 9) If the termination criterion is satisfied, go to step 10.
Step 10) Update the velocity and location of particle using
(20)–(22) and go to step 2. types of DG units are considered and compared with the pene-
Step 11) Output the optimal configuration of DG units and tration level with only inverter-based DG units. In the last case
Stop. study, the impact of customer owned DG unit installations on
the harmonic level and protection coordination is analyzed. It is
IV. RESULTS AND DISCUSSION worthy to note that variations in investment costs for DG instal-
The nonlinear optimization problem for maximizing DG pen- lation are not considered in this paper.
etration level was implemented in Matlab [31] and tested on the
A. System Under Study
IEEE-30 bus system. The constraint values on harmonic distor-
tion are estimated using the DHPF subroutine. The constraints The proposed problem formulation was tested on the 33-kV
on operating time and coordination time of OCR units are es- level IEEE-30 bus meshed distribution system as shown in
timated through conventional protection coordination analysis Fig. 5. The complete system data can be found in [36]. Total real
presented in Section II-B. Three case studies are conducted to and reactive power loads of the test system are 104.7 MW and
show the performance of the proposed optimization problem, 50.8 MVAr, respectively. In order to analyze the impact of DG
out of which the first two studies are carried on by considering penetration with various load levels, ten different load scenarios
utility owned DG units and the last case study focuses on cus- are considered along with the renewable energy availability
tomer owned DG units. In the first case, the maximum DG pen- as given in the Appendix. These scenarios are obtained using
etration level considering only inverter-based DG unit installa- fuzzy c-means clustering of the data given in United Kingdom
tion at all candidate buses are obtained. In the second case, both Generic Distribution System over a one-year period [37] and
1250
Fig. 5. Single line diagram of the IEEE 30-bus distribution system.
the same percentage is used for the above considered IEEE-30
bus distribution system. The total system rating
is 150 MVA. The distribution side power is fed through three
primary distribution transformers rated for 50 MVA, 10%
reactance, 132/33-kV level located at buses 2, 8, and 12. The
DG surplus power is fed back to the grid in cases where the
generation is more than the system demand. The directional
OCR are placed at both the ends of each feeder for protection.
There are 28 OCR units available in the distribution system and
initially all are coordinated. The number of candidate buses
selected for placing both type of DG is ten (which were chosen
randomly) and the locations include buses 2 to 10 and bus 12.
Three phase to ground fault at midway of each line denoted as
F15 to F30 is considered and the corresponding short circuit
currents and relay operating times are calculated.
All DG units are connected to the system through a step-up
transformer of 480 V/33 kV level with 5% reactance. The in-
verter-based DG units are considered as the source of harmonics
and used in the DHPF subroutine. The harmonic spectrum of
current injected by PWM inverter is considered to be the upper
bound of the individual harmonic current limits (only odd order
harmonics without the multiples of 3rd order harmonics) speci-
fied by IEEE-1547 standard [3] and also given in the Appendix.
The OCR unit has the inverse time current curves with the fol-
lowing constants 0.14 and 0.02 for A and B, respectively. The
coordination time interval (CTI) is assumed to be 0.3 s between
each backup and primary relay pair. The maximum operating
time for each relay is set to 2.5 s.
B. Case I
In this case, utility owned inverter-based DG units are con-
sidered to find the maximum DG penetration level. Since the in-
verter-based DG unit has minimal effect on the short circuit cur-
rent, the protection coordination constraints are not considered.
The capacity of inverter-based DG units is considered as the
parameter to be optimized for all ten candidate buses and also
specified in (23). An optimal DG penetration level of 26.94%
is obtained using the particle swarm optimization (PSO) algo-
rithm by placing six DG units at locations 4 to 9 with a ca-
pacity of 7.65 MW, 97.4 kW, 4.61 MW, 9.52 MW, 8.65 MW,
and 9.89 MW, respectively. The remaining part of the demand is
met by drawing power through three primary distribution trans-
formers. For example, for a peak load scenario, the real and re-
active power drawn are 65.8 MW and 36.2 MVAr, respectively.
IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 28, NO. 2, MAY 2013
Fig. 6. Convergence of the PSO algorithm.
Fig. 7. Minimum and maximum RMS voltage at each bus.
The convergence of particle swarm optimization algorithm is
shown in Fig. 6 where the population size is set to 20 and the
maximum number of iteration is set to 30. The RMS maximum
and minimum voltage obtained at each bus location out of ten
scenarios are shown in Fig. 7. It can be seen that the RMS max-
imum and minimum voltage obtained is 1.0768 p.u. at bus 8
and 0.9675 p.u. at bus 5. Similarly, the maximum and minimum
THD obtained at each bus location are presented in Fig. 8. The
maximum total harmonic distortion is 3.6879% at bus 8. It has
been noticed that the 5th order individual harmonic distortion at
bus number 8 is reaching the maximum limit of 3%. Thus the
DG penetration level is restricted by harmonic distortion from
inverter-based DG units.
C. Case II
In this case, the installation of utility owned both inverter-
based and synchronous-based DG units is considered in order to
estimate the maximum DG penetration level. As mentioned ear-
lier, the inverter-based DG unit causes harmonic distortion and
synchronous-based DG units affect short circuit current. Thus
voltage constraints, harmonic distortion constraints and the pro-
tection coordination constraints are considered in this case. The
capacity of both types of DG units is the parameter to be opti-
mized for all ten candidate buses and also specified in (23). In
order to estimate the protection coordination constraint viola-
tions, three phase to ground fault at locations F15 to F30 is con-
sidered and the corresponding short circuit currents and relay
operating times are calculated using (7).
PANDI et al.: DETERMINING OPTIMAL LOCATION AND SIZE OF DISTRIBUTED GENERATION RESOURCES 1251

TABLE II
OPTIMAL PRIMARY AND BACK-UP RELAY OPERATING TIMES

Fig. 8. Minimum and maximum THD at each bus.

TABLE I
OPTIMAL DG CONFIGURATION

The PSO method is applied to obtain the maximum DG pene-

tration level and the capacity of both types of DG units are given
in Table I. An optimal DG penetration level of 60.0059% is ob-
tained by placing one inverter type DG unit at location 2 and
also placing nine synchronous type DG units at locations 2 to
7, 9, 10, and 12. The penetration level from synchronous-based
DG units is 60.00% and from inverter-based DG units is only
0.059%. The inverter-based DG penetration level is almost zero
compared to the penetration level obtained in case I. It shows
that the effect of synchronous-based DG on protection coordi-
nation is less sensitive than the effect of inverter-based DG units
on harmonic distortions. It has been also noticed that the 5th
order individual harmonic distortion at bus number 8 is reaching
the maximum limit of 3% for the peak load scenario. The RMS
maximum and minimum voltage obtained among all the sce-
narios are 1.0749 p.u. at bus 8 and 0.9821 p.u. at bus 14. The
maximum total harmonic distortion among all ten scenario is
3.5838% at bus 11.
The relay operating time for all fault locations including pri-
mary and backup relay is shown in Table II. The minimum limits
on relay pick up current is set above the largest line flows
among all scenarios. It is worthy to mention, from Table II, that
the coordination constraint given in (17) is binding for few fault primary relay R6 and the backup relay R12 is 0.3 s for a fault at
locations. For example, the operating time difference between F17. The optimal relay settings obtained corresponding to the
1252
TABLE III
OPTIMAL RELAY IP AND TDS SETTINGS
TABLE IV
OPTIMAL DG PENETRATION FOR DIFFERENT MAXIMUM DG CAPACITY
maximum DG penetration level is given in Table III. In most
of the relay units, the time dial setting is bounded to the lower
value of 0.1. Major part of the demand is met by DG installation
and the remaining part of the demand is met by drawing power
through the three primary distribution transformers. Thus the
simultaneous consideration of both types of DG units increases
the penetration level by 33.06% which is 49.59 MW higher than
the penetration studies with only inverter type DG units.
Furthermore, the analysis was extended to consider various
DG capacity limits. The determination of the maximum pene-
tration level considering DG capacity limits of 1 MVA, 5 MVA,
and 10 MVA is presented in Table IV. The results show that,
for a DG capacity limit of 1 MVA, the optimal solution is to
install both types of DG at the maximum capacity (1 MVA)
with an overall penetration level of 13.33%. Thus for the ten
candidate locations, a 1-MVA inverter-based DG and 1-MVA
synchronous-based DG unit can be installed without violating
harmonic and protection limits. A total penetration level of
38.188% is obtained when considering 5-MVA DG capacity
limit. This is achieved by installing synchronous-based DG at
all candidate locations (at the maximum capacity of 5 MVA)
and inverter-based DG at five locations. Lastly, by considering
10-MVA DG capacity limit, the total penetration level is in-
creased to 60.059%. It can be noticed that as the DG capacity
limit increases, the penetration level of synchronous-based DG
increase. The penetration level of inverter-based DG depends
on the DG capacity limit as well as the contribution of the
synchronous-based DG units.
IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 28, NO. 2, MAY 2013
D. Case III
In a competitive electricity market environment, the end cus-
tomers are encouraged to install DG to meet their local demand
as well as to improve the performance of the system by selling
the surplus power available back to the network. The changes in
system performance and technical issues involved in connecting
the customer owned DG units with the distribution system is of
great importance to the utility [14]. Feasibility studies are con-
ducted by the utility to determine the technical impacts as well
as the cost associated with such impacts.
In this paper, feasibility studies of installing the customer
owned DG units of known type, size and location on voltage
violation, power quality and protection coordination are consid-
ered. The proposed optimization framework can aid utility plan-
ners in identifying any possible system violations as well as pos-
sible solutions. If the utility identifies violation in power quality
limits then the proposed algorithm can be used to determine the
size and parameters of the passive filter, using IEEE-1531 stan-
dard [38], that needs to be installed at the customer premises.
Another example is, if the utility identifies mismatches in pro-
tection coordination in the distribution system, modifications re-
quired for the protective relay setting can be determined. The
following step-by-step procedure explains the proposed feasi-
bility study conducted by the utility when they have a request
from end customer to install DG at certain bus location.
1) Perform both the DHPF and short circuit analysis without
connecting DG units, which gives the initial network
voltage, harmonic levels and optimal relay settings.
2) Customer request for installing “S” MVA DG at bus “i”.
3) Evaluate the values of voltages, harmonics and relay tim-
ings and check for constraint violations.
4) If any violations in harmonic limits are found because of
inverter-based DG installation, then design a passive filter
to maintain the power quality levels within the acceptable
limits.
5) If any violations in protection relay constraints are found
because of synchronous-based DG installation, then
evaluate the new optimal relay settings to maintain the
coordination.
Two different assessment study scenarios are provided in this
section to demonstrate the usefulness of the proposed method.
In the first scenario, several customers are requesting permission
to connect their DG units (size and location of DG were chosen
randomly) at their local premises as shown in Table V. The DG
units are connected as in the sequence shown in Table V. Con-
ducting feasibility assessment for as inverter-based DG of size 1
MVA at bus 2 results in no constraint violations. With a 1 MVA
inverter-based DG at bus 2 and the addition of another 1-MVA
synchronous-based DG at bus 5, the protection constraints are
violated. To accommodate a customer owned DG of size 1 MVA
(synchronous-based) at bus 5, the relay and TDS settings
would need to be modified at 34 out of 56 locations. Further
addition of a 2 MVA inverter-based DG at bus 8 and 5 MVA at
bus 3 will require no system modifications. Similarly, a 5-MVA
synchronous-based DG at bus 8 will result in protection coordi-
nation violation which can be mitigated by changing the relay
settings at 35 locations.
PANDI et al.: DETERMINING OPTIMAL LOCATION AND SIZE OF DISTRIBUTED GENERATION RESOURCES
TABLE V
ASSESSMENT STUDY OF CUSTOMER OWNED DG—SCENARIO 1
TABLE VI
ASSESSMENT STUDY OF CUSTOMER OWNED DG—SCENARIO 2
In the second scenario as shown in Table VI, after installing
customer owned DG at buses 2, 5, 12, and 6, the inverter-based
DG installation request at bus 3 violates 5th order individual
harmonic distortion limit at bus 8. In order to mitigate the har-
monic violation, the shunt passive filter design technique is con-
sidered along with the constraints on the voltage and current
limits in filter components which are set equal to 110% and
135%, respectively [38], [39]. The quality factor is fixed to 30
and the rating of the filter is varied in steps of 0.25 MVAr to re-
move the harmonic violation. It is found that a 5th order single
tuned filter (STF) of 0.5 MVAr rating located at bus 8 is suffi-
cient to remove the harmonic limit violation and leads to the fol-
lowing filter components values, capacitance F,
inductance mH, resistance Ohms.
It can be noticed that further installation of inverter-based DG
at bus 7 will result in no constraint violation. Thus, the proposed
optimization framework can help the utility planners in identi-
fying possible constraint violations as well as possible solutions
for customer owned DG installation. It is worthy to note that for
a larger meshed distribution system, there may be more interac-
tions between relays in terms of primary and backup. In such
case, adding a DG may have more impact on the relay settings
and could result in a larger amount of relay settings being mod-
ified to accommodate additional DG.
V. CONCLUSION
This paper introduces a new approach that can simultane-
ously determine the penetration level of utility owned inverter-
based and synchronous-based DG units to achieve maximum
DG penetration taking into account standard harmonic limits
and protection coordination constraints. The proposed method
is solved using the particle swarm optimization algorithm and
tested on the IEEE-30 bus looped distribution system. In the
first case, the results obtained for inverter-based DG penetra-
tion considering the constraints regulated by IEEE-519 stan-
1253
TABLE VII
INVERTER-BASED DG HARMONICS
TABLE VIII
LOAD AND RENEWABLE SOURCE SCENARIO
dard on harmonics is presented. In this case, the results show
that the maximum DG penetration level is limited by harmonic
distortion level. In the second case, both inverter-based and
synchronous-based DG units are considered simultaneously in
the optimization including harmonic and protection constraints.
The analysis shows that higher penetration level is achieved
when compared to the only inverter type DG penetration case. In
the last case, how the optimization framework feeds knowledge
to the utility planners to conduct feasibility assessment studies
for the approval of customer owned DG installation. The pro-
posed optimization method can serve as a good planning tool for
the utility operator to optimally allocate DG of different types
in distribution systems to achieve better penetration level.
APPENDIX
The current harmonic spectrum of the inverter-based DG
units is given in Table VII. The scenario of different load and
renewable source availability is given in Table VIII.
REFERENCES
[1] N. Jenkins, R. Allan, P. Crossley, D. Kirschen, and G. Strbac, Em-
bedded Generation, Institution of Engineering and Technology, 2008.
[2] M. H. Bollen and F. Hassan, Integration of Distributed Generation in
the Power System. New York: Wiley-IEEE Press, 2011.
[3] IEEE Standard for Interconnecting Distributed Resources With Elec-
tric Power Systems, IEEE Std. 1547-2003, 2009, vol. 15.
[4] H. Puttgen, P. MacGregor, and F. Lambert, “Distributed generation:
Semantic hype or the dawn of a new era?,” IEEE Power and Energy
Mag., vol. 1, no. 1, pp. 22–29, Jan.–Feb. 2003.
[5] C. Borges and D. Falcao, “Impact of distributed generation allocation
and sizing on reliability, losses and voltage profile,” in Proc. IEEE
Bologna Power Tech Conf., 2003, Jun. 2003, vol. 2, p. 5.
[6] J. Martinez and J. Martin-Arnedo, “Impact of distributed generation
on distribution protection and power quality,” in Proc. IEEE Power
Energy Society General Meeting, PES’09, Jul. 2009, pp. 1–6.
[7] R. Walling, R. Saint, R. Dugan, J. Burke, and L. Kojovic, “Summary of
distributed resources impact on power delivery systems,” IEEE Trans.
Power Del., vol. 23, no. 3, pp. 1636–1644, Jul. 2008.
[8] Smartconnect Use Case: D9—Utility Manages Utility-Owned Dis-
tributed Generation. [Online]. Available: asset.sce.com/microsite/Doc-
uments/ESC/D9_Use_Case_081223.pdf.
[9] A. Algarni and K. Bhattacharya, “Utility-owned DG units’ impacts
on distribution system operation,” in Proc. IEEE/PES Power Systems
Conf. Expo., PSCE’09, Mar. 2009, pp. 1–6.
1254
[10] A. Schwarzenegger, “Renewable distributed generation assessment:
Sacramento municipal utility district case study,” Utility Project
Report, 2005.
[11] R. Jabr and B. Pal, “Ordinal optimisation approach for locating and
sizing of distributed generation,” IET Gen., Transm., Distrib., vol. 3,
no. 8, pp. 713–723, Aug. 2009.
[12] F. Abu-Mouti and M. El-Hawary, “Optimal distributed generation al-
location and sizing in distribution systems via artificial bee colony al-
gorithm,” IEEE Trans. Power Del., vol. 26, no. 4, pp. 2090–2101, Oct.
2011.
[13] R. Dugan, T. McDermott, and G. Ball, “Planning for distributed gener-
ation,” IEEE Ind. Appl. Mag., vol. 7, no. 2, pp. 80–88, Mar./Apr. 2001.
[14] L. Luna and E. Parra, “Feasibility assessment of distributed genera-
tion interconnection,” in Proc. IEEE Trondheim Power Tech. 2011, Jun.
2011, pp. 1–7.
[15] J. Carrasco, L. Franquelo, J. Bialasiewicz, E. Galvan, R. Guisado, M.
Prats, J. Leon, and N. Moreno-Alfonso, “Power-electronic systems for
the grid integration of renewable energy sources: A survey,” IEEE
Trans. Ind. Electron., vol. 53, no. 4, pp. 1002–1016, Jun. 2006.
[16] IEEE Recommended Practices and Requirements for Harmonic Con-
trol in Electrical Power Systems, IEEE Std. 519-1992, 1993.
[17] R. Shayani and M. de Oliveira, “Photovoltaic generation penetration
limits in radial distribution systems,” IEEE Trans. Power Syst., vol.
26, no. 3, pp. 1625–1631, Aug. 2011.
[18] A. Bhowmik, A. Maitra, S. Halpin, and J. Schatz, “Determination of
allowable penetration levels of distributed generation resources based
on harmonic limit considerations,” IEEE Trans. Power Del., vol. 18,
no. 2, pp. 619–624, Apr. 2003.
[19] S. Brahma and A. Girgis, “Development of adaptive protection scheme
for distribution systems with high penetration of distributed genera-
tion,” IEEE Trans. Power Del., vol. 19, no. 1, pp. 56–63, Jan. 2004.
[20] N. Nimpitiwan, G. T. 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. 634–641, Jan. 2007.
[21] E. Coster, J. Myrzik, B. Kruimer, and W. Kling, “Integration issues of
distributed generation in distribution grids,” Proc. IEEE, vol. 99, no.
1, pp. 28–39, Jan. 2011.
[22] 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. 576–585, Apr.
2008.
[23] W. El-Khattam and T. Sidhu, “Resolving the impact of distributed re-
newable generation on directional overcurrent relay coordination: A
case study,” IET Renew. Power Gen., vol. 3, no. 4, pp. 415–425, Dec.
2009.
[24] S. Chaitusaney and A. Yokoyama, “Prevention of reliability degrada-
tion from recloser - fuse miscoordination due to distributed genera-
tion,” IEEE Trans. Power Del., vol. 23, no. 4, pp. 2545–2554, Oct.
2008.
[25] A. El-Zonkoly, “Optimal placement of multi-distributed generation
units including different load models using particle swarm optimi-
sation,” IET Gen., Transm., Distrib., vol. 5, no. 7, pp. 760–771, Jul.
2011.
[26] D. Q. Hung, N. Mithulananthan, and R. C. Bansal, “Analytical expres-
sions for DG allocation in primary distribution networks,” IEEE Trans.
Energy Convers., vol. 25, no. 3, pp. 814–820, Sep. 2010.
[27] S. Herraiz, L. Sainz, and J. Clua, “Review of harmonic load flow for-
mulations,” IEEE Trans. Power Del., vol. 18, no. 3, pp. 1079–1087,
Jul. 2003.
[28] E. F. Fuchs and M. A. S. Masoum, Power Quality in Power Systems
and Electrical Machines. New York: Academic/Elsevier, 2008.
[29] F. Pamplona and B. Souza, “Harmonic passive filter planning in ra-
dial distribution systems using genetic algorithms,” in Proc. IEEE/PES
Transmission and Distribution Conf. Expo.: Latin America, Nov. 2004,
pp. 126–131.
[30] Y.-Y. Hong and W.-F. Huang, “Interactive multiobjective passive filter
planning with fuzzy parameters in distribution systems using genetic
algorithms,” IEEE Trans. Power Del., vol. 18, no. 3, pp. 1043–1050,
Jul. 2003.
[31] Matlab, The Language of Technical Computing. [Online]. Available:
http://www.mathworks.com.
IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 28, NO. 2, MAY 2013
[32] M. A. S. Masoum, A. Jafarian, M. Ladjevardi, E. F. Fuchs, and W. M.
Grady, “Fuzzy approach for optimal placement and sizing of capacitor
banks in the presence of harmonics,” IEEE Trans. Power Del., vol. 19,
no. 2, pp. 822–829, Apr. 2004.
[33] J. Kennedy and R. Eberhart, “Particle swarm optimization,” in Proc.
IEEE Int. Conf. Neural Networks, 1995, Nov./Dec. 1995, vol. 4, pp.
1942–1948.
[34] Y. Shi and R. Eberhart, “A modified particle swarm optimizer,” in
Proc. 1998 IEEE Int. Conf. Evolutionary Computation Proceedings,
1998. IEEE World Congr. Computational Intelligence, May 1998, pp.
69–73.
[35] Eberhart and Y. Shi, “Particle swarm optimization: Developments, ap-
plications and resources,” in Proc. 2001 Congr. Evolutionary Compu-
tation, 2001, 2001, vol. 1, pp. 81–86.
[36] Power Systems Test Case Archive. [Online]. Available: http://www.ee.
washington.edu/research/pstca/.
[37] United Kingdom Generic Distribution System. [Online]. Available:
http://monaco.eee.strath.ac.uk/ukgds/.
[38] IEEE Guide for Application and Specification of Harmonic Filters,
IEEE Std. 1531-2003, 2003.
[39] G. W. Chang, H.-L. Wang, G.-S. Chuang, and S.-Y. Chu, “Passive har-
monic filter planning in a power system with considering probabilistic
constraints,” IEEE Trans. Power Del., vol. 24, no. 1, pp. 208–218, Jan.
2009.
V. Ravikumar Pandi (M’10) received the B.E. de-
gree from Madurai Kamaraj University, Tamil Nadu,
India, in 2003, the M.Tech. degree from Annamalai
University, Tamil Nadu, India, in 2005, and the Ph.D.
degree from the Indian Institute of Technology Delhi,
New Delhi, India, in 2010.
Currently he is working as a Post Doctoral Re-
searcher with the Masdar Institute of Science and
Technology, Abu Dhabi, United Arab Emirates. His
research interest includes power system optimiza-
tion, power quality, protection, transmission pricing,
renewable energy, artificial intelligence, and evolutionary algorithms.
H. H. Zeineldin (M’06) received the B.Sc. and M.Sc.
degrees in electrical engineering from Cairo Univer-
sity, Cairo, Egypt, in 1999 and 2002, respectively.
In 2006, he received the Ph.D. degree in electrical
and computer engineering from the University of Wa-
terloo, Waterloo, ON, Canada.
He worked for Smith and Andersen Electrical En-
gineering Inc. where he was involved with projects
involving distribution system design, protection,
and distributed generation. He then worked as a
Visiting Professor at the Massachusetts Institute of
Technology (MIT), Cambridge. Currently, he is an Associate Professor with
the Masdar Institute of Science and Technology in Abu Dhabi, United Arab
Emirates. His current interests include power system protection, distributed
generation, and deregulation.
Weidong Xiao (M’07) received the M.Sc. and Ph.D.
degrees from the University of British Columbia,
Vancouver, BC, Canada.
He is an Assistant Professor of Electrical Power
Engineering at Masdar Institute of Science and
Technology. Before his academic career, he worked
with MSR Innovations Incorporation as a R&D
engineering manager focusing on projects related
to integration, monitoring, evaluation, optimization,
and design of photovoltaic power systems. His re-
search area includes power electronics, photovoltaic
power systems, digital control techniques, and industrial applications.