See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/4078049

Impact of distributed generation on Volt/Var control in distribution networks

Conference Paper · July 2003

DOI: 10.1109/PTC.2003.1304390 · Source: IEEE Xplore

CITATIONS READS

139 1,401

3 authors, including:

Taher Niknam Ali Reza Shirani

Shiraz University of Technology Monenco Iran consulting enngieers company
367 PUBLICATIONS 19,814 CITATIONS 26 PUBLICATIONS 814 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Taher Niknam on 02 June 2014.

The user has requested enhancement of the downloaded file.

 Paper accepted for presentation at 2003 IEEE Bologna Power Tech Conference, June 23th-26th, Bologna, Italy

Impact of Distributed Generation on

Volt/Var Control in Distribution Networks
T. Niknam A.M. Ranjbar A.R. Shirani

concluded that this figure could be as high as 30%[1].

Abstract— As power system in many countries is going to be Apart from advantages of DGs, they can affect some
restructured and deregulated. After deregulation, because of parameters of power systems such as stability, voltage control,
numerous advantageous of Distributed Generation (DG), the power quality and so on where in this paper the impact of
number of this kind of generators are going to be increased. DGs these kind of generators on Volt/Var control is going to be
can affect entire system and especially distribution networks. studied.
One of the important control schemes at distribution system that Volt/Var control is one of the important control schemes at a
DGs can change it, is Volt/Var control. This paper presents an distribution substation, which conventionally involves
efficient approach for Volt/Var control in radial distribution
regulation of voltage, and reactive power at substation bus.
networks takes DGs performance into consideration. In general
The control is achieved by Load Tap Changer (LTC), Voltage
Distributed Generations can be considered as PV or PQ nodes. In
this paper DGs are modeled as PV nodes. The goal of this Regulators (VR) and Capacitors. Some Volt/Var control
approach is to minimize power losses at distribution system algorithms have already been developed by researchers [2-6].
through controlling the tap of Load Tap Changer (LTC), size of Since DGs can change Volt/Var control scheme at distribution
substation capacitor, local controller settings and voltage network, this paper presents a new approach for Volt/Var
amplitude of DGs. DGs, Voltage Regulators, Local Controllers, control in radial distribution system takes DGs performance
and Load Tap Changer (LTC) are modeled completely and the into consideration. In general Distributed generations can be
optimization problem has been solved by using Genetic considered as PV or PQ nodes. In this paper DGs are modeled
Algorithm. Finally the method is tested on IEEE 34 bus radial as PV nodes where these performance are combined with
distribution feeders.
controlling of Load Tap Changer (LTC), substation capacitor,
Keyword: Distributed Generation, PV node, Volt/Var Control,
Genetic Algorithm. local controller settings to minimize power losses at
distribution system. In this effort Genetic algorithm is used to
I. INTRODUCTION minimize objective function and then it is implemented to
verify and testify on IEEE 34 bus radial distribution feeders.

D
ISTRIBUTED Generation (DG) is defined as the
generation which is dispersed throughout a utility’s II. OBJECTIVE FUNCTION
service territory and either connected to the utility’s
distribution system or small and Grid Isolated. DG From a mathematical standpoint the Volt/Var control
technologies include Photovoltaic, Wind Turbines, Fuel Cells, optimization problem is a minimization problem with
Micro Turbines and so on. inequality constraints. The objective function is summation of
During, the last decade many electrical power systems around losses in distribution system for load variation. The value of
the world have been deregulated and DG is predicted to play the objective function is determined through power flow
an increasing role in the electric power system of the near solution.
future. A study by the Electric Power Research Institute The objective function is given by equation (1).
(EPRI) indicates that by 2010, 25% of the new generation will
be distributed, and also, a study by Natural Gas Foundation F ( x ) = min Ploss dt (1)

A.M. Ranjbar, is with Niroo Research Institute and Sharif University of

technology (Email: [email protected]).
Subject to:
A.R. Shirani is with Niroo Research Institute (Email:[email protected]). 1. Vi min ≤ Vi ≤ Vi max i = 1,2,..,N
T. Niknam is with Niroo Research Institute, Sharif University of technology
and Engineering Dept. Science and Research Branch, Azad University 2. Pf min ≤ Pf ≤ Pf max
(Email: [email protected])
Niroo Research Institute, End of Pounak Bakhtari Blvd., Shahrak Ghods, 3. Qgi min ≤ Qgi ≤ Qgi max i = 1,2,...,Ng
Tehran, Iran.
Sharif University of Technology, Azadi Ave., PO Box 11 365-93 63, Tehran, 4. Ti min ≤ Ti ≤ Ti max i =1,2,...,Nt
Iran.
Azad University, Engineering Dept. Science and research Branch Azad
University, PO Box 193 95-66 63, Tehran , Iran.

0-7803-7967-5/03/$17.00 ©2003 IEEE

5. Qci min ≤ Qci ≤ Qci max i = 1,2,...,Nc In general voltage of DGs can be controlled in two ways:
A) Balanced three-phase voltage control of DG which use
6. Load Flow Equations g ( P, Q, V , δ ) = 0 . Synchronous Generators and Balanced three phase
Where: inverters such as Gas Turbines, small hydro and
Vi = bus voltage, geothermal plants. In this case the incremental relation
Vimin , Vimax= minimum and maximum voltage for each bus , between the magnitude of voltage and magnitude of
Pf = Power factor in substation, reactive current injection is expressed as:
Pfmin, Pfmax= minimum & maximum power factor in substation,
Ploss = sum of losses in line and transformers, [Z ][I Injec ] = [∆V ]
Qgi = reactive power for each generator,
Qgimin ,Qgimax = minimum and maximum reactive power for
[I ] = [I , I ,..., I
Injec Injec1 Injec 2 InjecN ] = [Z ] [∆V ]
T −1
(2)
each generator, [I ] = [Z ] [∆V ]
Injec
−1

Ti = tap for LTC and VR,

Timin, Timax = minimum and maximum tap for each [∆V ] = [∆V1 , ∆V2 ,..., ∆VN ]T = VSet − VCalc
transformer,
Qci= reactive power for each capacitor, Where:
Qcimin , QCimax = minimum and maximum reactive power for [Z]N*N: Sensitivity Matrix, which is calculated, based on
each capacitor, positive sequence impedance of lines.
dt = duration of time for load , [∆V]: positive sequence voltage magnitude mismatch
Nt= number of transformers, vector for PV nodes.
Nc = number of capacitors, [IInjec]: injection of reactive current vector in PV nodes.
Ng = number of generators. |Vset| : Scheduled voltage magnitude vector for PV nodes.
|VCalc|: Calculated positive sequence voltage magnitude
III. UNBALANCED THREE PHASE LOAD FLOW vector for PV nodes.
In three phase unbalanced power flow, the following N: Number of DGs
components are modeled by their equivalent circuits in term of
inductance, capacitance, resistance and injected current. B) Unbalanced three phased voltage control of DGs
which connected to network through an inverter that its
a) Distributed Generators: DGs are modeled as PV node. phase voltage could be controlled separately such as Fuel
b) Transformers: transformers are modeled as equivalent cell, Micro turbine and so on. In this case the incremental
circuit with virtual current injections. relation between the magnitude of voltage and magnitude
c) Capacitors: Capacitors are represented by their of reactive current injection is expressed as:
equivalent injected currents.
d) Demands or Loads: system loads are basically [Z ][I Injec ] = [∆V ]
considered asymmetric loads; because of single load [I ] = [I
Injec Injeca1 , I Injecb1 , I Injecc1 ,..., I InjecaN , I InjecbN , I InjeccN ]
T

and unequal three phase loads.

= [Z ] [∆V ] (3)
−1
In this paper a network-topology-based on three-phase
distribution power flow algorithm has been used. Two [I ]
Injec = [Z ] [∆V ]
−1

matrices are used to obtain the power flow solution. They are
the Bus Injection to Branch Current (BIBC) and the Branch [∆V ] = [∆Va1, ∆Vb1, ∆Vc1,..., ∆VaN , ∆VbN , ∆VcN ]T
Current to Bus Voltage (BCBV) matrices [7]. = VSet − VCalc
Where:
IV. DISTRIBUTED GENERATION MODEL [Z]3N*3N: Sensitivity Matrix which is calculated based on
Depending on the contract and control status of a generator, it three phase impedance matrix of lines.
may be operated in one of the following modes: [∆V]: voltage magnitude mismatch vector for PV nodes.
1. In “parallel operation” with the feeder, i.e., the [IInjec]: injected reactive current vector in PV nodes.
generator is located near and designed to supply a large |Vset| : Scheduled voltage magnitude vector for PV nodes.
load with fixed real and reactive power output. The net |VCalc|: Calculated voltage magnitude vector for PV nodes.
effect is the reduced load at a particular location.
2. To output power at a specific power factor. Since in power system there is nonlinear relation between
3. To output power at a specific terminal voltage. voltage and current, we have used an iterative method for
The generation nodes in the first two cases can be well voltage mismatch correction as follows:
represented as PQ nodes. The generation nodes in the third Step1: solve the power flow equation with initial condition
case must be modeled as a PV node. In this paper, generators (Qgi=0).
are modeled as PV nodes.
PV nodes are modeled in similar manner as presented in [7]. Setp2: calculate mismatched voltage for each PV nodes. If
these values are less than thresholds go to next step, otherwise 2(a). In this case correspondent matrix should be refactorized.
stop. Another way to simulate the tap positions change is to modify
model Fig.l(b) by adding virtual injected currents as it is
Step3: calculate injection reactive current amplitudes. shown in Fig. 2(b).

In case A): Balanced three phase voltage control

- γ
γ
γ
I Injeca = I Injec e j( + 90+δ via )
i
- γ
γ
γ
I Injecb = I Injec e j( + 90+δ vib ) (4)
i
- γ
Fig. 2. (a) Transformer model with a new ratio. (b) Equivalent with virtual
γ
γ
I Injecc = I Injec e j( + 90 +δ vic ) current injections.
i
In order to make voltages and currents in the systems 2a and
2b to be the same, fictitious injection currents are calculated as
In case B): Unbalanced three phase voltage control follows:
IM = (1−(t + ∆t))*Y *VN
- γ
γ
γ
I Injeca = I Injeca e j( + 90+δ via ) (6)
i
γ γ γ
j( -+ 90 +δ vib ) IN = (1−(t + ∆t))*Y *VM + ((t + ∆t)2 −1)*Y *VN
I Injecb = I Injecb e (5)
i
γ γ γ
j( -+ 90 +δ vic ) VI. GENETIC ALGORITHM
I Injecc = I Injecc e
i
Genetic Algorithm is searching and optimization method
In above equations positive/negative signs are used when based on a model of evolution adaptation in nature. It is very
voltage mismatch is positive/negative. powerful search algorithm and is different from convention
search algorithms. GA does not need derivatives or other
Step4: calculate reactive power for each DGs based on auxiliary knowledge.
obtained voltage and current values. GA works with a population of individuals and each
individual stands for a solution. The quality of a solution is
Step5: check calculated reactive power with DGs reactive evaluated by its fitness, which is calculated by fitness function
power limits. If any calculated reactive power is outside the [8].
bound discard the extra value, and consider this PV node as In this paper, an Integer string instead of binary coding are
PQ node, and then update sensitivity matrix. used to represent value of variables, and includes these
process:
Step6: calculate injected reactive current based on calculated 1. Representation and initialization
reactive power in previous step. 2. Fitness function
3. Reproduction Operation
4. Crossover Operation
Step7: Power flow calculation, go to step2.
5. Mutation Operation
V. VOLTAGE REGULATOR AND LTC MODEL
VII. FLOW CHART OF ALGORITHM
Voltage regulators and LTCs often control voltage along
Since Volt/Var control is an optimization problem, Genetic
primary feeders. Voltage Regulators are autotransformers with
Algorithm has been used to solve it. Fig .3. shows flow chart
individual taps on their windings. LTC and Voltage of algorithm.
Regulators are modeled as follows: At first in this approach, initial population is produced based
Assume the transformer has connected between buses M and on control variables including voltage amplitude of DGs,
N, and has initial tap ratio (t) and physical admittance (Y) reactive power of substation capacitor, tap of LTC and local
[Fig.1 (a)]. This transformer is described by the π model, with controller setting (for this paper there are 7 control variables).
indirect representation of the transformer tap ratio [Fig. 1(b)] Initial population has to meet constraints. The value of taps
by its series and shunt admittances. and capacitor reactive power is considered discrete. Then for
each member of initial population, with considering of local
controllers and PV nodes, unbalanced three-phase power flow
is solved. After that, electric power losses for each member
are calculated and sorted, then a number of good members
(Ngood) that have minimum losses, are selected. New offspring,
based on selected population, are produced by Roulette wheel
reproduction rule. Any new offspring outside the bound is
Fig. 1. Transformer and Voltage Regulator model discarded in favor of the other two. Mutation operator is
When the transformer tap ratio is changed from t to t+∆t the applied to each gene according to mutation probability
transformer model should be changed as it is shown in Fig. independently. After mutation, losses are calculated and sorted
for each member of new population. Then a number of good
members that have minimum losses are selected. This process It is assumed, there are 3 generators, that they have been
is repeated, until convergence is met. located on 9, 23 ad 27 buses. The characteristic of generators
is given in Table.1.
Read Data, Define Control Parameters
TABLE I
CHARACTERISTIC OF GENERATORS
Create Initial Population G1 G2 G3
Active Power (kW) 90 120 150
Max Reactive Power( kVar) 72 96 120
Solve 3-Phase Radial Power Flow with Min Reactive Power (kVar) -54 -72 -90
Considering Modeling of Local Controllers and PV Maximum Voltage 1.015 1 1
nodes Minimum Voltage 0.98 0.98 0.98
Kind of voltage control Case A Case A Case B

Figure 5 shows the load curve profile in all load point in the
Calculate Power Losses
network.

Select Number of Good Population

Error < e
Yes
No Fig.5. Variation of load.
Stop Now for using the GA to determine the state variable of the
system i.e. tap of LTC, size of capacitors, tap of Voltage
regulators and voltage amplitude of generators, the following
Mating and Reproduction assumption is made:
Initial population: 2000
Mutation Number of good population: 300
Number of load level: 3
Limit of voltage magnitude: 0.95-1.05
Solve 3-Phase Radial Power Flow with RTVR1< RTVR2 <RTCap1 <RTCap2
Considering Modeling of Local Controllers and PV Limit of power factor in substation: 0.95-1
nodes
Limit of tap position: .97-1.05
Size of tap: 0.0001
Limit of substation capacitor: 0-1000 kVar
Calculate Power Losses Limit of local capacitors:
C1:0-450 kVar
Fig3. Flow chart of algorithm C2:0-300 kVar
In this approach, local controllers are coordinated based on Mutation: µ = 0.04 e − Counter
theirs response times. In the other word, at first local ε=0.000001
controllers in which have the shortest response time, will Ngood
response to power flow alterations. Slower controller response Error = ∑n =1
Cost n − Cost Ngood +1
will be based on output of faster local controllers. It is
assumed response time of Voltage Regulators is less than Where:
capacitors. Counter and Costi are number of iteration and value of
objective function for the ith population respectively. RTi is
VIII. SIMULATION response time for each local controller.
In the following section the application of the method for
In this section, Volt/Var control algorithm is tested on the 34 various load conditions are presented.
Bus IEEE test feeder, whose one line diagram is given in
Fig.4. The feeder line and load data is taken from [9]. A) PEAK LOAD
In this case it is assumed that in all load nodes, the load is in
its maximum level. Table II represents result of simulation for
this case.

Fig.4. 34 Bus IEEE test feeder with DGs

 TABLE II TABLE IV
OUTPUT RESULTS OF SIMULATION FOR PEAK LOAD RESULT OF SIMULATION FOR 80% PEAK LOAD
With DG Without DG With DG Without DG
Tap of LTC 1.03 1.03 Tap of LTC 1.03 1.03
Size of substation capacitor (kVar) 522 858 Size of substation capacitor (kVar) 318 654
PF in substation 0.9995 0.9978 PF in substation 0.9936 0.9987
Power losses (kW) 36.61735 87.416 Power losses (kW) 16.9689 54.47581
Tap of Voltage Regulator (1) 1.0068 1.03 Tap of Voltage Regulator (1) 0.9997 1.0209
Tap of Voltage Regulator (2) 1.0135 1.0173 Tap of Voltage Regulator (2) 1.0022 1.0106
Size of Capacitor1 (kVar) 0 0 Size of Capacitor1 (kVar) 0 0
Size of Capacitor2 (kVar) 0 0 Size of Capacitor2 (kVar) 0 0
QG1 (kVar) 72 - QG1 (kVar) 72 -
QG2 (kVar) 96 -
QG2 (kVar) 96 -
QG3A (kVar) 40 -
QG3A (kVar) 40 -
QG3B (kVar) 12.408 -
QG3B (kVar) 26 -
QG3C (kVar) 40 -
QG3 C(kVar) 40 -
Voltage amplitude of DG1 1.00909 -
Voltage amplitude of DG1 1.00913 -
Voltage amplitude of DG2 0.9941 -
Voltage amplitude of DG2 0.99324 -
Voltage amplitude of DG3 0.9993 -
Voltage amplitude of DG3 0.999969 -
The voltage profile also has been presented in Table III.
TABLE III The voltage profile has been presented in Table V.
VOLTAGE PROFILE FOR PEAK LOAD TABLE V
No Va Vb Vc No Va Vb Vc VOLTAGE PROFILE FOR 80% PEAK LOAD
1 1.0300 1.0300 1.0300 1 1.0300 1.0300 1.0300 No Va Vb Vc No Va Vb Vc
2 1.0299 1.0299 1.0299 2 1.0298 1.0299 1.0298 1 1.0300 1.0300 1.0300 1 1.0300 1.0300 1.0300
3 1.0296 1.0296 1.0297 3 1.0293 1.0294 1.0294 2 1.0299 1.0299 1.0299 2 1.0298 1.0299 1.0299
4 1.0241 1.0247 1.0251 4 1.0200 1.0216 1.0211 3 1.0297 1.0298 1.0298 3 1.0294 1.0295 1.0295
5 1.0246 5 1.0215 4 1.0260 1.0267 1.0269 4 1.0222 1.0234 1.0231
6 1.0177 1.0192 1.0197 6 1.0094 1.0127 1.0116 5 1.0267 5 1.0234
7 1.0126 1.0148 1.0155 7 1.0009 1.0056 1.0040 6 1.0218 1.0234 1.0235 6 1.0138 1.0164 1.0156
8 1.0067 1.0105 1.0116 8 1.0091 1.0174 1.0145 7 1.0184 1.0208 1.0208 7 1.0072 1.0109 1.0097
9 1.0066 1.0105 1.0116 9 1.0090 1.0173 1.0144 8 1.0096 1.0137 1.0136 8 1.0113 1.0177 1.0155
10 1.0065 10 1.0089 9 1.0095 1.0136 1.0136 9 1.0112 1.0176 1.0155
11 1.0027 11 1.0052 10 1.0094 10 1.0111
12 1.0019 12 1.0043 11 1.0065 11 1.0082
13 1.0052 1.0087 1.0099 13 1.0068 1.0149 1.0119 12 1.0058 12 1.0075
14 1.0087 14 1.0149 13 1.0086 1.0125 1.0124 13 1.0095 1.0157 1.0135
15 1.0051 1.0086 1.0097 15 1.0066 1.0148 1.0117 14 1.0124 14 1.0157
16 1.0023 1.0053 1.0063 16 1.0022 1.0103 1.0067 15 1.0085 1.0124 1.0123 15 1.0093 1.0156 1.0133
17 1.0023 1.0053 1.0062 17 1.0021 1.0102 1.0066 16 1.0066 1.0102 1.0099 16 1.0058 1.0121 1.0094
18 1.0052 18 1.0102 17 1.0066 1.0102 1.0099 17 1.0058 1.0120 1.0093
19 0.9973 0.9995 1.0002 19 0.9942 1.0023 0.9976 18 1.0102 18 1.0120
20 1.0004 1.0011 1.0011 20 0.9950 1.0033 0.9963 19 1.0033 1.0064 1.0056 19 0.9995 1.0058 1.0022
21 0.9864 0.9870 0.9871 21 0.9809 0.9893 0.9821
20 0.9985 1.0009 0.9992 20 0.9972 1.0035 0.9982
22 0.9861 0.9868 0.9868 22 0.9806 0.9890 0.9819
21 0.9873 0.9897 0.9880 21 0.9860 0.9923 0.9869
23 0.9999 1.0005 1.0004 23 0.9941 1.0024 0.9952
22 0.9871 0.9895 0.9878 22 0.9858 0.9921 0.9867
24 0.9999 24 0.9941
23 0.9982 1.0005 0.9987 23 0.9965 1.0027 0.9973
25 0.9991 0.9995 0.9994 25 0.9930 1.0013 0.9940
24 0.9982 24 0.9965
26 0.9991 0.9995 0.9994 26 0.9930 1.0012 0.9939
25 0.9976 0.9998 0.9980 25 0.9956 1.0019 0.9963
27 0.9990 0.9993 0.9993 27 0.9928 1.0010 0.9937
28 0.9989 0.9992 0.9992 28 0.9928 1.0009 0.9937 26 0.9975 0.9998 0.9980 26 0.9956 1.0018 0.9963
29 0.9989 0.9992 0.9992 29 0.9927 1.0009 0.9937 27 0.9975 0.9996 0.9979 27 0.9955 1.0017 0.9962
30 0.9990 0.9994 0.9993 30 0.9929 1.0012 0.9938 28 0.9974 0.9996 0.9978 28 0.9954 1.0016 0.9961
31 0.9989 0.9994 0.9992 31 0.9928 1.0011 0.9938 29 0.9974 0.9995 0.9978 29 0.9954 1.0016 0.9961
32 0.9989 0.9994 0.9992 32 0.9928 1.0011 0.9938 30 0.9975 0.9997 0.9979 30 0.9955 1.0018 0.9962
33 0.9989 33 0.9928 31 0.9974 0.9997 0.9978 31 0.9955 1.0018 0.9962
34 0.9989 0.9994 0.9992 34 0.9928 1.0011 0.9938 32 0.9974 0.9997 0.9978 32 0.9955 1.0018 0.9962
With DG Without DG 33 0.9974 33 0.9954
Bus is single phase 34 0.9974 0.9997 0.9978 34 0.9955 1.0018 0.9962
With DG Without DG
Bus is single phase
B) 80% PEAK LOAD C)60% PEAK LOAD
In the second case it is assumed that the load in various nodes The final case is that one which assumed the loads in various
are 0.8 of its peak value. Table IV represents result of nodes as 0.6 of its peak value. Table VI represents result of
simulation for this case. simulation for 60%peak load.
 TABLE VI GA optimization algorithm in Volt/Var control in distribution
RESULT OF SIMULATION FOR 60% PEAK LOAD
network incorporating DGs and control devices.
With DG Without DG
Tap of LTC 1.02 1.03 In tables II, IV and VI, comparison of system losses between
Size of substation capacitor (kVar) 246 390
pre and post installation of DGs is shown. After, installation
PF in substation 0.9697 0.9993
of DGs, since they reduce the line current flow, the system
Power losses (kW) 5.59806 29.92189
losses in each load level is reduced. For example, for 60%
Tap of Voltage Regulator (1) 0.9942 1.0107
peak load, DGs are caused power losses in distribution system
Tap of Voltage Regulator (2) 0.9923 1.0033
decrease from 29.92189kW to 5.59806 kW. As can be seen
Size of Capacitor1 (kVar) 0 0
Distributed generation placement affects on active and
Size of Capacitor2 (kVar) 0 0
reactive power flows, so if they are placed at suitable
QG1 (kVar) 52.4436 -
locations, power losses can be decreased enormously.
Also, the results of these tables show that in peak and %80
QG2 (kVar) 96 -
peak load cases, only phase B of third DG could be considered
QG3A (kVar) 37.9446 -
as PV node in Volt/Var control process. On the other hand, it
QG3B (kVar) 20.9672 -
is necessary that these DGs be operated in theirs maximum
QG3 C(kVar) 36.8281 -
reactive powers to obtain minimum losses in these cases.
Voltage amplitude of DG1 1.01049 -
Comparison of voltage profile between pre and post
Voltage amplitude of DG2 0.99243 -
installation of DGs for load variation has been shown in tables
Voltage amplitude of DG3 0.99704 -
III, V and VII. The results show that imbalance has been
The voltage profile for this case has been presented in Table reduced when we have used DGs.
VII.
TABLE VII
VOLTAGE PROFILE FOR 60% PEAK LOAD
No Va Vb Vc No Va Vb Vc
IX. CONCLUSION
1 1.0300 1.0300 1.0300 1 1.0300 1.0300 1.0300 Since number of DGs is increasing and also DGs affect
2 1.0300 1.0300 1.0300 2 1.0298 1.0299 1.0298 voltage and reactive power control, it is necessary, that impact
3 1.0299 1.0299 1.0299 3 1.0295 1.0296 1.0296 of DGs on Volt/Var control be studied. This paper presented
4 1.0278 1.0281 1.0284 4 1.0242 1.0251 1.0249
an efficient algorithm for Vol/Var control in distribution with
5 1.0281 5 1.0251
6 1.0254 1.0262 1.0265 6 1.0181 1.0200 1.0194 DGs. In the three phase unbalanced power flow calculation
7 1.0235 1.0247 1.0251 7 1.0132 1.0159 1.0151 while modeling the local controllers of devices, the
8 1.0127 1.0149 1.0155 8 1.0117 1.0163 1.0148 distribution system components are modeled by their
9 1.0127 1.0149 1.0155 9 1.0116 1.0162 1.0148 equivalent circuits in terms of inductance, capacitance,
10 1.0126 10 1.0116 resistance and injected current. DGs can be modeled as PV or
11 1.0104 11 1.0093 PQ nodes. In this paper whilst DGs were modeled as PV
12 1.0099 12 1.0088
13 1.0122 1.0141 1.0148 13 1.0103 1.0149 1.0133
nodes, an approach was presented to model of them. Genetic
14 1.0141 14 1.0148 algorithm is used to obtain the solution of the optimization
15 1.0121 1.0141 1.0148 15 1.0102 1.0147 1.0132 problem. By using this algorithm the performance of IEEE 34
16 1.0111 1.0127 1.0133 16 1.0076 1.0121 1.0102 bus radial test feeders when DGs existed or not existed was
17 1.0111 1.0127 1.0133 17 1.0076 1.0121 1.0102 analyzed and it has been shown that while the system losses
18 1.0127 18 1.0121 decreased enormously in the case of existence of DGs,
19 1.0093 1.0104 1.0108 19 1.0029 1.0074 1.0049
imbalance of voltages have improved. So it is revealed that by
20 0.9977 0.9979 0.9979 20 0.9967 1.0012 0.9974
21 0.9893 0.9895 0.9895 21 0.9883 0.9928 0.9890
proper placement of DGs and using appropriate controller for
22 0.9891 0.9893 0.9893 22 0.9882 0.9927 0.9888 them, it is possible to have much better control for Volt/Var in
23 0.9975 0.9976 0.9976 23 0.9962 1.0006 0.9968 network along with decreasing the system losses in network.
24 0.9975 24 0.9962
25 0.99709 0.99712 0.99709 25 0.9955 1.0000 0.9960 X. REFERENCE
26 0.99708 0.99711 0.99708 26 0.9955 0.9999 0.9960
27 0.99703 0.99703 0.99703 27 0.9954 0.9998 0.9959 [1]. T. Ackermann, G. Andersson and L. Soder, “ Distributed
28 0.99699 0.99696 0.99698 28 0.9954 0.9998 0.9958 Generation : A Definition” Elsevier Science, Dec. 2000,
29 0.99699 0.99695 0.99697 29 0.9954 0.9998 0.9958 P.P.195-204.
30 0.99702 0.99708 0.99700 30 0.9955 0.9999 0.9959 [2]. K. H. Kim and S.K. You, “Voltage Profile Improvement
31 0.99697 0.99705 0.99697 31 0.9954 0.9999 0.9959 by Capacitor Placement and Control in Unbalanced
32 0.99697 0.99705 0.99697 32 0.9954 0.9999 0.9959
Distribution Systems Using GA”, IEEE 1999, p.p. 800-805.
33 0.99694 33 0.9954
34 0.99697 0.99704 0.99697 34 0.9954 0.9999 0.9959
[3]. M. E. Baran and M. Y. Hsu, “Volt/Var Control at
With DG Without DG Distribution Substations”, IEEE Trans. On Power Systems,
Bus is single phase vol.14, No.1, Feb. 1999, p.p. 312-318.
[4]. I. Roytelman and V. Ganesan, “Coordinated Local and
Comparison between results achieved by using GA in the Centralized Control in Distribution Management Systems”,
above mentioned problem and those found by other IEEE Trans. On Power Delivery, vol.15, No.2, April 2000,
investigations [3] shows the high accuracy and applicability of p.p. 718-724.
[5].V. Borozan, M.E. Baran and D. Novosel, “Integrated
Volt/Var Control in Distribution Systems” IEEE 2001,
p.p.1485-1490.
[6]. Roytelman and V. Ganesan, “Modeling of Local
Controllers in Distribution Network Applications”, IEEE
Trans. On Power Delivery, vol. 15, No.4, Oct. 2000,
p.p.1232-1237.
[7]. T. Niknam and A.M. Ranjbar, “ Impact of Distributed
Generation on Distribution Load Flow”, International Power
System Conference Tehran, Iran, Oct. 2002.
[8]. R. L. Haupt, S.E. Haupt, “Practical Genetic Algorithm”,
John Wiely & Sons Inc., 1999.
[9]. W. H. Kersting, “Radial Distribution Test Feeders”, IEEE
Trans. On Power Systems, Vol.6, No.3, Aug. 1991.

XI. BIOGRAPHIES
Taher Niknam was born in Shiraz in Iran. He received his
B.S and M.S degree from Shiraz University and Sharif
University of Technology respectively. He is Ph.D student at
Sharif University of Technology and a faculty at the Power
System Department of South Tehran Branch of Azad
University. His interests include power system restructuring,
impact of DGs on power system and power electronics. He is
working in Niroo Research Institute and Iran Electricity
Generation and Transmission Company (TAVANIR) in
Tehran.

Ali Mohammad Ranjbar received the M.S. and Ph.D.

degrees in Electrical Engineering from Tehran University in
1967 and Imperial College of London University in 1975,
respectively. Since then, he has been at Sharif University of
Technology, Department of Electrical Engineering, where he
is currently a full professor. Dr. Ranjbar is the Editor-in-Chief
of Journal of Electrical Science and Technology since 1989
and the Director of Niroo Research Institute since 1996 too.
His main research interests are in the areas of electric power
systems protection and operation, and Electrical Machine.

Ali Reza Shirani was born in Isfahan, Iran in 1965. He has

been graduated in electrical engineering from Sharif
University of technology Tehran, Iran in 1988. He is the Vice
President of Niroo Research Institute (NRI). His main
research interest is power system analysis and techniques.

View publication stats