Improvement of Low Frequency Oscillation Damping by Allocation and

Electrical Power and Energy Systems 52 (2013) 207220
Contents lists available at SciVerse ScienceDirect
Electrical Power and Energy Systems

journal homepage: www.elsevier.com/locate/ijepes
Improvement of low frequency oscillation damping by allocation and

design of power system stabilizers in the multi-machine power system
A.R. Fereidouni a, B. Vahidi a,, T. Hoseini Mehr b, M. Tahmasbi c
a
Department of Electrical Engineering, Amirkabir University of Technology, Tehran 15916-34311, Iran

Department of Electrical and Electronics Engineering, Karaj Branch, Islamic Azad University, Karaj, Iran
c
Department of Electrical Engineering, Neka Branch, Islamic Azad University, Neka, Iran
b
a r t i c l e
i n f o
Article history:
Received 3 September 2011
Received in revised form 4 July 2012
Accepted 28 March 2013
Available online 24 April 2013
Keywords:
Power system stability
Low-Frequency Oscillation (LFO)
Participation Factor (PF)
Eigenvalue analysis
a b s t r a c t
Some of the earliest power system stability problems included spontaneous power system oscillations at
low frequencies. These Low-Frequency-Oscillations (LFOs) are related to the small-signal stability of a
power system and are harmful to obtain the maximum power transfer. A contemporary solution to this
problem is the addition of Power System Stabilizers (PSSs) to the automatic voltage regulators on the
generators in the power system. In this paper, allocation of PSSs in an interconnected power system with
inter-area modes has been determined by eigenvalue analysis, and PSSs for the allocated generators have
been designed by a frequency response method. Furthermore, for designing PSSs by the frequency
method, this paper proposed a new linear power system model which can consider both local and
inter-area oscillations of the power network. Designed PSSs based on the proposed model improved
damping performance of PSSs which have been designed by the Single-Machine-Innite-Bus (SMIB)
model. The stabilization performance of the designed PSSs (by the proposed approach and linear power
system model) on the LFO modes have been veried in two multi-machine power system standard models (IEEE 9-bus and 14-bus).
2013 Elsevier Ltd. All rights reserved.
1. Introduction
Two of the most important design criteria for multi-machine
power systems are transient stability and damping of electromechanical modes of sustained oscillation [16]. Stability of power
systems is one of the most important aspects in power electrical
system operation. This is because the power system must maintain
frequency and voltage levels in the desired level, under any disturbance, such as a sudden increase in the load, loss of one generator
or switching out of a transmission line, during a fault [7].
With the increase of the scale and complexity of the interconnected power networks, the problems on the various potential
power oscillations, which have the signicant impact on the system stability and security operation, have been drawn more and
more attention [810].
Power system oscillations were rst reported in northern
American power network in 1964 during a trial interconnection
of the Northwest Power Pool and the Southwest Power Pool [11].
Two types of oscillation phenomena can occur on the present
power system. One is where the oscillation of one generator at a
Corresponding author. Tel.: +98 21 64543330; fax: +98 21 66406469.
E-mail addresses: fereidouni_aut@aut.ac.ir (A.R. Fereidouni), vahidi@aut.ac.ir
(B. Vahidi), hoseini.mehr124@gmail.com (T. Hoseini Mehr), m2mehdi@yahoo.com
(M. Tahmasbi).
0142-0615/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijepes.2013.03.030
specic power plant has an inuence on the system. This type of

oscillation is called local-mode oscillation and its behavior is
mainly limited to the local area in the vicinity of the power plant,
and it seldom inuences the rest of the system. It has been known
that the local oscillation is likely to occur when power is transmitted over long-distance transmission lines from a power plant at a
remote location. This type of system can be accurately modeled
using the Single-Machine-Innite-Bus (SMIB) system model [12].
The other case has been known as inter-area mode oscillation. This
is the case where the low-frequency oscillation is maintained between sets of generators in an interconnected power system. The
simplest type of low-frequency oscillation in the inter-area mode
is between two interconnected areas. The inter-area mode oscillation has a long history [7]. As a classic oscillation mode, there are
relative mature technologies and devices such as kinds of power
system stabilizers equipped as a part of the additional excitation
system of machine unit to provide the efcient damping ratio to
suppress the local oscillation. So far, PSS has been used as an effective and economical facility to tackle the problem [13]. In [14], it
has been shown that the appropriate selection of PSS parameters
can achieve satisfactory performance during system upsets.
The use of PSS in power system has been both economical and
successful in improving the power system stability, and is expected
to be installed on many generators connected to the system. However, there are different kinds of power plants connected to the
208
A.R. Fereidouni et al. / Electrical Power and Energy Systems 52 (2013) 207220
Nomenclature
Sn
Vn
fn
xl
ra
Xd
X 0d
X 00d
T 0do
T 00do
Xq
X 0q
X 00q
T 0qo
T 00qo
M
D
d
power rate
voltage rate
frequency rate
leakage reactance
armature resistance
d-axis synchronous reactance
d-axis transient reactance
d-axis sub-transient reactance
d-axis open circuit transient time constant
d-axis open circuit sub-transient time constant
q-axis synchronous reactance
q-axis transient reactance
q-axis sub-transient reactance
q-axis open circuit transient time constant
q-axis open circuit sub-transient time constant
inertia constant
damping coefcient
rotor angle
power system, such as fossil fuel, hydro and nuclear power plants
in which generators have different characteristics. In addition,
there are pumped storage power plants in actual application.
Whether a PSS is to be installed or not depends, in part, on the type
of power generation. If low-frequency oscillation is damped by
installing an appropriate number of control devices at appropriate
locations within the power system, even further economic benets
can be expected. For this reason, it is very important to have a
method for determining the locations of PSSs in a realistic power
system model. To improve the power system stability of the entire
system, a smaller number of PSSs has been designed and installed
in a real-size system having inter-area mode oscillations [15].
In this paper, allocation of PSSs has been performed by using an
eigenvalue analysis on the system, so that the dominant generator
with the greatest inuence on both the power system stability, and
the low-frequency oscillation becomes the candidate for PSS installation. The proposed approach utilized a PSS for this dominant generator with the capability for damping the system mode. Moreover,
in this paper, a new model based on linear model of the dominant
generator has been proposed for designing PSS in which both local
and inter-area mode oscillations of the power system have been
considered. This means that after nding the dominant generator
in the power system for supplying the PSS via eigenvalues analysis,
the PSS has been designed based on the proposed linear model of
the dominant generator regarding the oscillation modes of power
system (local and inter-area modes). The proposed linear model
is called Developed-Single-Machine-Innite-Bus (DSMIB) system.
As a result, this paper has been shown that damping performance
of the designed PSS based on the proposed approach and proposed
power system model has been improved in comparison with the
designed PSS based on the SMIB model. In the application of the
proposed method, the paper utilized the standard IEEE 9 bus system [16] for investigating the proposed approach in details and
IEEE 14 bus system for validating the proposed approachy [17].
2. Eigenvalue analysis of multi-machine power system
In an analysis of the system stability, eigenvalues of a power
system model have been derived and evaluated. Through analyzing
eigenvalues, characteristics of system dynamic states are understood without a time domain simulation. Hence, the eigenvalues
rotor speed
q-axis transient voltage
d-axis transient voltage
q-axis sub-transient voltage
d-axis sub-transient voltage
mechanical power
eld voltage
reference value of the generator eld voltage
stabilizing signal
innite bus voltage
conductance
active current
active power
susceptance
reactive current
reactive power
e1q
e1d
e01q1q
e02d2d
Pm
Vf
Vref
UPSS
Vb
g
Ip
Pn
b
IQ
Qn
analysis is efcient in appraising the system stability for a

multi-machine power system model [18,19]. The system eigenvalues have been evaluated with respect to the components of
the power grid; that is to say, the eigenvalues with regard to electrical distances between generators. Therefore, the power system
stability has been evaluated in the multi-machine power system
with respect to the network formations. A situation, in which all
eigenvalues are in the negative real half of the complex plane,
has been specied for a stable system. In addition, an eigenvalue
existing in the vicinity of the imaginary axis, impact the system
stability signicantly. Furthermore, the imaginary parts of system
eigenvalues dominate the system oscillation frequency in the time
domain.
2.1. Small signal modeling
The act of a normal power system can be described by a set of
rst order nonlinear ordinary differential equations and a group
of nonlinear algebraic equations. It can be written in the following
form by using vector-matrix notation:
x_ f x; u;
2 3
y1
6y 7
6 2 7
7
y6
6 .. 7;
4. 5
y gx; u
2 3
g1
6g 7
6 2 7
7
g6
6 .. 7
4. 5
ym
gm
where x is the vector of state variables, such as rotor angle and

speed of generators. The column vector y is the vector of outputs,
and g is a vector of nonlinear functions relating state and input variables to output variables. Although power system is nonlinear, it
can be linearized by small signal stability at a certain operating
point. Suppose that x0 and u0 are the equilibrium points of this
power system, then based on direct feedback, it can be expressed
as the following standard form [20]:
Dx_ ADx BDu;
Dy C Dx DDu
where Dx, Dy, Du, A, B, C and D are the state vector of dimension n,
the output vector of dimension m, the input vector of dimension r,
the state or plant matrix of size n n, the control or input matrix of
size n r, the output matrix of size m n, and the (feed forward)
matrix that denes the proportion of input which appears directly

in the output, size m r respectively.
2.2. Participation factor
If ki is an eigenvalue of A, vi and wi are non-zero column and row
vectors respectively so that the following relations hold:
Av i k i v i ;
i 1; 2; . . . ; n
wi A ki wi ;
i 1; 2; . . . ; n
where the vectors vi and wi, are known as right and left eigenvectors
of matrix A and they are henceforth considered normalized so that:
wi v i 1
P P1 P2 P n
with
P 1i
mode, and vice versa. Since Uki measures the activity of xk in the ith
mode and Wik weighs the contribution of this activity to the mode,
the product Pki measures the net participation. The effect of multiplying the elements of the left and the right eigenvectors is also to
make Pki dimensionless (i.e., independent of the choice of units).
Considering the eigenvector normalization, the sum of the particiP
pation factors associated with any mode ni1 Pki or with any state
Pn
variable k1 Pki is equal to 1. In these analyses, the participation
factor is more efcient than normal eigenvectors, and useful for the
eigenvalue analysis. In this paper, the state variable inuencing,
the dominant root is determined by the evaluation of the participation factor; in addition, a dominant generator that should be supplied with a PSS, is chosen.
One problem in using right and left eigenvectors individually for

identifying the relationship between the states and the modes is
that the elements of the eigenvectors are dependent on units and
scaling associated with the state variables. As a solution to this
problem, a matrix called the participation matrix (P), which combines the right and left eigenvectors is proposed in [21] as a measure of the association between the state variables and the modes.
209
U1i Wi1
7 6
7
6
6 P 2i 7 6 U2i Wi2 7
7 6
7
6
Pi 6 . 7 6 . 7
6 .. 7 6 .. 7
5 4
5
4
Uni Win
P ni
3. Case studies
First, a four-machine test power system (the IEEE 9-bus standard system) (Fig. 1) is used in this study for comprehending
how the proposed method perform in details. Secondly, a ve-machine test power system (the IEEE 9-bus standard system) (Fig. 2)
is used for validating the proposed method in a larger power
system.
First model contains four machines on each side of a transmission network. The ve loads on the buses 5, 6, 7, 8 and 9 (Load 1, 2,
3, 4 and 5) are modeled as constant active and reactive power loads
and the ve other loads on the Buses 5, 6, 7, 8 and 9 (Load 10 , 20 , 30 ,
40 and 50 ) are dened as follows:
P gV 2 Ip V Pn
2
Q bV IQ V Q n
where Uki (vki), and Wik (wik) are the elements on the kth row and
ith column of the modal matrix U (kth entry of the right eigenvector
Ui) and the element on the ith row and kth column of the modal
matrix W (kth entry of the left eigenvector Wi) respectively.
The element is termed the participation factor [21]. It is a measure of the relative participation of the kth state variable in the ith
9
10
The ParkConcordia model is used for the equations of the synchronous generators. The generators are modeled by the sixth order
model. The sixth order model is acquired assuming the presence
of a eld circuit and an additional circuit along the d-axis and two
additional circuits along the q-axis. The ith generator has six state
variables (d; x; e0q ; e0d ; e00q and e00d ) and the following equations:
Fig. 1. IEEE 9-bus test system.
210
d_ i xri xbi
x_ ri Pmi Pei Di xri 1=Mi
e_ 0qi
e_ 0di
e_ 00qi
e_ 00di
e0qi
xdi
x0di
11
T 00 x00di
doi
xdi x0di idi
T 0doi x0di
e0di xqi x0qi
T 00qoi x00qi
T 0qoi x0qi
e00qi e0qi x0di x00di

e00di e0di x0qi x00qi
12
!,
T 0doi
13
T 0qoi
14
!,
xqi x0qi iqi
T 00doi x00di
xdi x0di idi
T 0doi x0di
T 00qoi x00qi
T 0qoi x0qi
!,
T 00doi
15
T 00qoi
16
!,
xqi x0qi iqi
where the electrical power output of the ith generator is as follows:
Pei
n
X
Ei Ej Y ij coshij di dj
17
j1
The governor and AVR of the generators G1 and G4 are dened as

type 1 and type 2 respectively. Furthermore, both the governor
and AVR of the generators G2 and G3 are dened as type 1 (Fig. 3).
mode oscillations. Local problems may also be associated with

oscillations between the rotors of a few generators close to each
other. Such oscillations are called inter-machine or interplant
mode oscillations. Usually, the local plant mode and interplant
mode oscillations have frequencies in the range of 0.72.0 Hz [8].
Global small-signal stability problems are caused by interactions
among large groups of generators and have widespread effects.
They involve oscillations of a group of generators in one area
swinging against a group of generators in another area. Such oscillations are called inter-area mode oscillations [8].
Large interconnected systems usually have two distinct forms
of inter-area oscillations:
A very low frequency mode involving all the generators in the
system. The system is essentially split into two parts, with generators in one part swinging against machines in the other part.
The frequency of this mode of oscillation is on the order of 0.1
0.3 Hz.
Higher frequency modes involving modes subgroups of generators swinging against each other. The frequency of these oscillations is typically in the range of 0.40.7 Hz.
In this paper, frequencies between 0.4 and 0.7 and between 0.7
and 2 Hz for global modes and local modes are considered,
respectively.
4. Allocation of power system stabilizers

This paper intends to improve the stability versus the low-frequency oscillation via installing PSSs into the power system. The
allocated and designed PSSs are investigated in the IEEE 9-bus
and 14-bus standard system. In this paper, the peak load condition,
where the generators operate at their output margin, is used for
allocation of PSSs.
4.1. Characteristics of small-signal stability problems
In the large power systems, small-signal stability problems may
be either local or global (inter-area) in nature. Local problems involve a small part of the system. They may be associated with rotor
angle oscillations of a single generator or a single plant against the
rest of the power system. Such oscillations are called local plant
4.2. Eigenvalue analysis of the rst case study

In this paper, the eigenvalue analysis has been implemented to
the power system mode. Before fullling eigenvalue analysis, it is
necessary to calculate power ow. For this reason, power ow calculation has been carried out. The results of power ow are shown
in Table 1. In order to obtain eigenvalues, the dominant eigenvalues are shown in Fig. 4. The eigenvalue analysis takes notice
of conjugate eigenvalues; hence, the eigenvalues that take the positive numbers in the imaginary part are numbered in Fig. 4. A slant
black line area and a slant blue line area in the gure shows that
the eigenvalues in these areas have a peculiar oscillation between
(a)
G4
G1
G5
(b)
Breaking for 4 cycles
G2
(c)
G3
Fig. 2. IEEE 14-bus test system.
Fig. 3. Models of the governor and AVR. (a) AVR type 1, (b) AVR type 2, (c) governor
type 1 [16].
211
Table 1
Power ow results for the IEEE 9-bus standard test system.
Bus
Unit
V
(p.u.)
Phase
(rad)
Pgen
(p.u.)
Qgen
(p.u.)
Pload
(p.u.)
Qload
(p.u.)
Bus1
Bus2
Bus3
Bus4
Bus5
Bus6
Bus7
Bus8
Bus9
1.03
1.05
1.032
1.02
1.011764
1.039684
1.020901
1.024705
1.000477
0.168967
0.13577
0.328133
0.015025
0.090199
0.063157
0.12216
0.25658
0.05384
9.318182
0.272727
0.859114
11.16001
3.6E15
1.39E17
3.3E16
2.8E16
3.6E15
2.501564
0.047204
0.116297
3.549358
6E15
6.28E16
1.4E16
2.2E15
4.44E16
0
0
0
0
9.090909
0.113636
0.625
0.397727
11.36364
0
0
0
0
1.818182
0.022727
0.06687
0.045455
2.727273
Table 2
Dominant oscillation modes.
Mode
Mode
no.
Eigenvalue
Oscillation frequency
(Hz)
Inter-area
mode
Mode 6
0.1096+j4.0258
0.640724472
Mode 7
0.0252+j4.1093
0.654026611
Mode 8
Mode 9
0.1087+j6.2086
0.7872+j10.4862
0.988130252
1.668931436
Local mode
Fig. 4. System dominant eigenvalues in IEEE 9-bus test system.
2.5 and 4.4 (rad/s) (inter-area mode oscillations) and 4.4 and 12.5
(rad/s) (local mode oscillations) respectively. The result of the
eigenvalues analysis indicated that Mode 6 and Mode 7 are the
low frequency dominant inter-area modes and Mode 8 and Mode
9 are the low frequency dominant local modes in the IEEE 9-bus
standard system. When these modes have been stabilized, the
power system stability is improved. The eigenvalues and oscillation frequencies of the four dominant modes are listed in Table 2.
Fig. 5. Linearized model of the SMIB system [15].
4.3. Determination of the dominant generators for installing PSSs by

utilizing participation factor
Participation Factors
0.9
Block diagrams of damping constant and synchronizing coefcients of a generator and an AVR with a PSS are shown in Fig. 5.
This gure shows that load angle variation (Dd) operates as the
synchronous torque for the electrical oscillation of the generator,
in addition angular velocity (Dx) delaying 90 in phase the load
angle variation signal, functions as the damping torque. Therefore,
each torque of the generator is expressed and analyzed in the design and evaluation of PSS parameters. The constants K0 and D0 in
the gure, denote the variations in damping torque and synchronous torque due to equipping with both a PSS and an AVR. The
modes dominating the low-frequency oscillation have been obtained in the eigenvalues analysis.
In this section, an approach for dening dominant generators
for supplying PSSs into them is introduced. Therefore, the load angle of a generator that has the highest participation factor on the
main dominant inter-area modes, affects the power system stability. Then PSS needs to be installed into this generator. In addition,
the angular velocity deviation of a generator that has the highest
participation factor on the main dominant inter-area modes, affects the low-frequency-oscillation damping. Then, PSS needs to
be installed into this generator too. The selected generators are
candidates for supplying and designing PSSs in them.
In the rst case study, Modes 6 and 7 are dominant inter-area
modes (Fig. 4, Table 2), in which Mode 6 has a little lower
frequency than Mode 7 (almost the same frequency). Therefore,
in this case, Mode 7 (because it is nearer to the imaginary axis than
0.8
0.7
0.8
0.6
0.6
0.5
0.4
0.4
0.2
0.3
0
50
50 0.2
40
40
30
Eigenvalues
30
20
0.1
20
10
10
State Variables
Fig. 6. System participation factors in IEEE 9-bus for all eigenvalues and state
variables.
Mode 6) is selected for the main dominant inter-area mode. In this

paper, the participation factor has been utilized to determine the
dominant generator. Participation factors for the IEEE 9-bus standard system are shown in Fig. 6. Fig. 7 shows the participation factors of the angular velocity deviation and load angle deviation of all
generators on the dominant inter-area modes. As seen in this gure, G3 has the highest participation factor (load angle deviation
and angular velocity deviation) on Mode 7 (Mode 6 too) among
other generators. Therefore, the results of evaluating the participation factor show that G3 dominates the low-frequency-oscillation
and inuences the power system stability.
212
Fig. 7. System participation factors in IEEE 9-bus for inter-area dominant mode oscillations.
Fig. 9. The Single-machine-innite-bus system.
Fig. 8. Block diagram of the linearized model of the multi-machine power system.
5. Designing PSS based on linear model with focus on the

dominant generator by considering both local and inter-area
modes
After dening the dominant generator, in this section, a PSS
based on a new linear model of the dominant generator is designed. As mentioned earlier, in the rst case study, four oscillation
modes exist that two of them are inter-area modes and another
two oscillations are local modes. In general, the SMIB is used for
designing PSSs but in this model only local mode oscillations are
considered (regular design). Hence, in this paper, a new linear
model is proposed including both local and inter-area modes.
Fig. 10. Block diagram of the linearized model of the SMIB system.
5.1. Power system models for studying low-frequency oscillations

It would be desirable to create a linear model to analyze every
equipment in the power system accurately. However, power systems are a huge network with multiple generators and transformers which are connected via transmission lines. For this reason, it is
difcult to create a precise model with all details. The depth of the
system analysis depends on the size of the power system in question, and the accuracy of the model must be compromised to some
degree. Local mode oscillations have been known to occur at the
generator connected to the system by tie-lines that are weak for
the case of the innite bus model. The analysis of local mode oscil-
Fig. 11. The Developed-Single-machine-normal-bus system
lations is performed using a detailed model based on the SMIB system model. However, the inter-area mode occurs in a very large
system of interconnected, multi-machine systems; therefore, it is
difcult to perform a detailed analysis [15].
PSSs are very effective controllers in enhancement of the damping of LFOs, since PSS can increase damping torque for inter-area
modes by introducing extra signals into the excitation controllers
213
The SMIB system is a simple model but in this model, inter-area

mode oscillations are not considered. In practical applications,
there are different kinds of power plants connected to the power
system. Hence, LFOs have different frequencies. For this reason,
while PSSs is being designed, it is necessary to consider the effects
of the rest of the power system. As a result, it is essential to have a
model as simple as the SMIB system and as exact as the multi-machine power system (Fig. 8) for designing PSSs. For this purpose,
this paper proposed a new linear model of the power system that
includes the designed PSSs with considering inter-area and local
modes. This model is called the Developed Single Machine Innite
Bus (DSMIB).
Fig. 12. Phase diagram of the ith generator.
already provided with generators. PSSs are designed by linearized

models of the power system. The SMIB system is used for designing
of PSSs that comprises a generator, a transmission line and an innite-bus which has been substituted the rest of the power system.
5.2. Developed Single Machine Innite Bus Systems (DSMIB)

To have a precise and complete linear model of the power system, it is necessary to model all the power system components
Fig. 13. Block diagram of the DSMIB system.
Table 3
Applied oscillations to the linear model of G3 for designing PSS3 (DSMIB model).
Mode
x (rad/s)
DV8
Dh 8
Mode 7
Mode 8
Mode 9
4.0258
6.2086
10.4862
0.1sin(4t) 0.01sin(6t) + 10sin(10t)
0.01sin(4t)
0.124705 0.711 t
0.03742 + 1.072 t
Inter-area and local modes
The variation of the magnitude and phase of the voltage of the bus 8
(a)
(b)
Fig. 14. The variation of (a) the magnitude (pu) and (b) phase (rad) of the voltage of the bus 8 in the IEEE 9-bus power system
214
Table 4
The Optimized parameters of the PSS3 (Fig. 16).
Parameter
T1 = T3
T2 = T4
DSMIB
SMIB
4.33
8.32
3.60
8
91.47
55.42
5.28
x 10
-3
Fitness
5.26
5.24
5.22
5.2
Fig. 15. Determination of the parameters of the PSS3 by the linear model of G3.
5.18
5.16
20
40
60
80
100
Number of Iteration
Initial
Parameter
Start
Finish
Yes
Satisfied Stopping
Rule?
Get a Parameter
Position
No
Change Models
Configuration
Penalty Factor
Update the
Parameters
Do
Optimization
Update the Best

Fitness
Perform
Simulation
Yes
No
Check the system

condition rules
Feasible Mode?
Results of
simulation
Optimization Package
Simulation Model
Fig. 16. Optimization method on stochastic simulation.
x 10
Fig. 18. Convergence of PSO for the SMIB model.
-3
point of view, the magnitude and angle of the bus that has been attached to the 3rd generator via transmission lines oscillate to the
effects of power system oscillations. Therefore, if in the SMIB system changes of these two variables are considered yields the
DSMIB system so that any impacts of the generators are not ignored. Fig. 9 displays a power system which represents the SMIB
system. The innite bus represents the thevenin equivalent of a
large interconnected power system. The nonlinear equations that
describe the generator and excitation system have been represented as follows:
d_ x0 x
1
x_ Pm P e Dm x
M
1
xd xe 0 xd x0d
E_ 0q 0 EFD 0
E
V b cosd
xd xe q x0d xe
T do
1
E_ FD K A Eref K A V t EFD
TA
18
19
20
21
Fitness
The above equations can be linearized for small oscillation

around an operating condition [9,22,23] and be cast in the block
diagram that is depicted in Fig. 10. Consider the 3rd generator that
is shown in Fig. 1. The power, inductive voltage and voltage of the
bus 8 are dened as follows (Figs. 11 and 12) (following equation
expressed as general, thus ith is used instead of 3rd):
5.8
5.6
5.4
20
40
60
80
100
Number of Iteration
Pgi V j E0qi =x0di sindi hj 1=xqi 1=x0di V 2j =2 sin 2di hj

22
Fig. 17. Convergence of PSO for the DSMIB model.
(like the multi-machine power system) or the impacts of the rest of

the power systems by one or several inputs. Consider the multimachine power system, as depicted in Fig. 8, the impacts of different generators on the ith generator are applied by K1ij to K6ij(i j),
but for example as displayed in Fig. 1 all effects and counteracts
transmit to the 3rd generator (the dominant generator) via transmission line 8-3 that has been connected to the power network.
It means that these impacts can be considered with adding them
to the bus 8 which is connected to the 3rd generator. Therefore,
the impacts of the rest of the power system can be modeled by
measuring variables of the transmission line 8-3 like the active
and reactive powers or the magnitude and angle of the voltage of
the bus 8. In the SMIB system the magnitude and angle of the innite bus is supposed to be constant. Furthermore, from a practical
h
i
Eqi xdi =x0di E0qi V j cosdi hj V j cosdi hi
23
V i V j jxe Iqi jIdi ejdi
24
where

Idi E0qi V j cosdi hj =x0di
25
Iqi V j Sindi hj =xqi
26
Moreover if Eqs. (22)(24) are linearized around a certain operating

point results:
0
DPgi k1 Ddi k2 DE0qi k1 Dhj k2 DV j
27
0
0
DEi k4 Ddi 1=k3 DE0qi k4 Dhj k3 DV j
0
0
DV i k5 Ddi k6 DE0qi k5 Dhj k6 DV j
28
29
215
SMIB
DSMIB
0.03
0.02
0.01
0
0
Angular Velocity Deviation (pu)
Load Angle Deviation (rad)
1.5
x 10
-3
SMIB
DSMIB
1
0.5
0
-0.5
-1
0.5
1.5
Time (s)
0.15
SMIB
DSMIB
0.1
0.05
0
-0.05
0.5
1.5
2.5
(b)
2.5
3.5
Terminal Voltage Devation (pu)
Internal Voltage Deviation (pu)
(a)
-0.1
Time (s)
0.06
SMIB
DSMIB
0.04
0.02
0
-0.02
-0.04
0.5
1.5
Time (s)
Time (s)
(c)
(d)
2.5
3.5
Fig. 19. The dynamic responses (a) load angle deviation (rad), (b) angular velocity deviation (pu), (c) internal voltage deviation, and (d) terminal voltage deviation when just
the mechanical power have been changed (20% for four cycles).
Therefore, if Eqs. (22)(24) are substituted into Eqs. (18)(21) and

then acquired equations are linearized, block diagram of the DSMIB
is obtained (Fig. 13). Comparison between Figs. 11 and 13 approves
that DVj and Dhj substituted the rest of the power system which are
not modeled in the SMIB system. It means that any kinds of oscillations with different frequencies can be applied with these inputs
and then PSSs can be designed regarding the all impacts of the
power system. To acquire the constants of the DSMIB system Eqs.
(22)(24) should be linearized around a certain operating point.
Therefore, constants K 1 ; K 01 ; K 2 ; K 02 ; K 3 ; K 03 ; K 4 ; K 5 ; K 05 ; K 6 and K 06 are
as follows:

K 1 V j E0qi =x0di cosdi hj 1=xqi 1=x0di V 2
j cos 2di hj
Dd_ i x0 Dxi
42

1
0
0
0
_i
K 1 Ddi K 1 Dhj K 2 DEqi K 2 DV j DPmi Dmi Dxi
Dx
M
43

1
1
44
DE_ 0qi 0
K 4 Ddi K 04 Dhj DE0qi K 03 DV j DEFDi
K3
T doi
DE_ FDi

1
KA
DEFDi
K 5 Ddi K 05 Dhj K 6 DE0qi K 06 DV j DErefi U PSS
TA
TA
45
5.3. Designing PSS
30
K 01 K 1
31
K2
V j =x0di sindi
K 02
E0qi =x0di sindi
K3
K 03
hj
hj
32
1=xqi
x0di =xdi
1=x0di V j
sin 2di
hj
33
34
hj
35
K 4 xdi x0di =x0di V j sindi hj
36
xdi
x0di =x0di cosdi
37
K 04 K 4
0

V qi
x
V di xqi
di
V sindi hj
V cosdi hj
K5
V i
x0di j
V i
xqi j
38
K 05 K 5
39

V qi xe
40
K6
V i x0di
0

V x

V qi
x
qi

di
K 06
di
cos
d

h
sin
d

h
41

i
j
i
j
Vi
x0di
Vi
xqi
The linearized model of the DSMIB system that has been linearized
around the certain operating point is given as follow:
In this section, after proposing the DSMIB model for designing

parameters of PSSs in the power system, it is time to design the
parameters of the PSS of the dominant generator in the IEEE 9bus standard system (G3-Fig. 1) with the DSMIB model. After that,
the parameters of the PSS3 are design by the SMIB model again to
analyze and compare the results (load angle deviation, angular
velocity deviation, terminal voltage deviation, and internal voltage
deviation) with the results which have been obtained by the
DSMIB model.
5.3.1. Designing the PSS3 based on DSMIB model (proposed model)
As mentioned earlier, G3 have been chosen for supplying a PSS
(PSS3) in the IEEE 9-bus standard system. Therefore, in this section,
the PSS3 is designed based on the proposed model (DSMIB).
There are four dominant modes that Modes 6 and 7 are interarea and Modes 8 and 9 are local. As shown in Table 2, the frequencies of the Modes 6, 7, 8 and 9 are 0.640724472, 0.654026611,
0.988130252 and 1.668931436. Therefore, the PSS3 is designed
with considering these frequencies (Table 3) and the variation of
the magnitude and phase of the voltage of the bus 8 (Fig. 14) in
the DSMIB model.
The block diagram of the PSS3 which is used for G3 is shown in
Fig. 15. In this study, the PSS-based controller parameters are
SMIB
DSMIB
0.5
0.4
0.3
0.2
0.1
0
-0.1
0.5
1.5
2.5
3.5
SMIB
DSMIB
-0.6
1.5
2
0
-2
0
0.5
1.5
(b)
-0.4
SMIB
DSMIB
(a)
-0.2
0.5
-3
Time (s)
x 10
Time (s)
2.5
3.5
Terminal Voltage Deviation (pu)
Internal Voltage Deviation (pu)
Angular Velocity Deviation (pu)
Load Angle Deviation (rad)
216
2.5
3.5
2.5
3.5
0.05
0
-0.05
-0.1
-0.15
-0.2
0.5
1.5
Time (s)
Time (s)
(c)
(d)
Fig. 20. The dynamic responses (a) load angle deviation (rad), (b) angular velocity deviation (pu), (c) internal voltage deviation, and (d) terminal voltage deviation when in
addition to changing the mechanical power (20% for four cycles), DV3, and Dh3 have been applied to the two models.
1.3
0.6
Syn 1,DSMIB-With PSS3 Proposed Design
0.5
Syn 2, None
-With PSS3 Regular Design
Syn 1,SMIB
0.4
0.3
Load Angle (rad)
Load Angle (rad)
Syn 1-None
1.2
Syn 2, SMIB
Syn 2, DSMIB-With PSS3 Proposed Design
1.1
1
0.2
0.1
10
15
20
25
30
35
0.9
40
10
15
Time (s)
(a)
Syn 3, None
1.6
Syn 3 , DSMIB-With PSS3 Proposed Design
1.2
1
10
15
20
30
35
40
Syn 4-None
25
30
35
Syn 3, SMIB
1.4
0.8
25
(b)
40
Load Angle (rad)
Load Angle (rad)
1.8
20
Time (s)
Syn 4, SMIB
Syn 4, DSMIB-With PSS3 Proposed Design
0.8
0.6
0.4
10
15
Time (s)
20
25
30
35
40
Time (s)
(c)
(d)
Fig. 21. Load angle with the PSS3 and without the PSS3 in the IEEE 9-bus. (a) generator 1, (b) generator 2, (c) generator 3 and (d) generator 4.
optimized. PSO (PSO algorithm has been explained in Appendix A)

has been employed to search for the optimal PSS3 controller
parameters. The objective function that has been used in this paper
is demonstrated in Eq. (46). This robust controller can be obtained
by solving Eq. (46). Where J1 = dx3, J2 = dv3 and c = tsim which tsim,
dx3, dv3, and k are the simulation time, angular velocity deviation
of G3, terminal voltage deviation of G3, and weighting parameter
(in this case k1 = 10 and k2 = 1) respectively. Therefore, the design
problem can be formulated as the following optimization problem
(Fig. 15):
Minimize tness subject to:
0:1 T min
< T 1 T 3 < T max
10
1
1
0:02 T min
< T 2 T 4 < T max
10
2
2
1k
fitness
min
<k<k
max
100
c k1 J1 k2 J2 dc
0
46
217
Syn 2-None
Syn 1-None

Syn 1SMIB
Syn 1,DSMIB-With PSS3 Proposed Desin
1.0005
0.9995
10
15
20
25
30
35
Angular Velocity (pu)
1.001
1.001
1.0005
40
10
15
Time (s)
20
Syn 3
-None
Syn 3,SMIB
1.001
-With PSS3 Proposed Design
Syn 3-DSMIB
1.0005
1
0.9995
0.999
0.9985
30
35
40
(b)
10
15
Angular velocity (pu)
25
Time (s)
(a)
1.0015
Syn 2, SMIB
Syn 4
1.0002
1
0.9998
0.9996
20
Syn 4-None
1.0004
10
15
20
Time (s)
Time (s)
(c)
(d)
25
30
35
40
Fig. 22. Angular velocity with the PSS3 and without the PSS3 in the IEEE 9-bus. (a) generator 1, (b) generator 2, (c) generator 3 and (d) generator 4.
pmSyn 2-None
pmSyn 1-None
pmSyn 1,SMIB-With PSS3 Regular Design
9.36
pmSyn 1-With PSS3 Proposed Design
9.355
9.35
9.345
9.34
0
10
15
20
25
30
35
Mechanical Power (pu)
9.365
0.2732
0.2731
0.273
40
pmSyn 2,DSMIB-With PSS3 Proposed Design
0.2733
10
15
20
Time (s)
Time (s)
(a)
(b)
11.19
0.862
0.8615
0.861
0.8605
0.86
pmSyn 3-None
0.8595
0.859
10
15
20
25
30
25
30
35
40
pmSyn 4-None
11.18
11.17
11.16
11.15
10
15
20
Time (s)
Time (s)
(c)
(d)
25
30
35
40
Fig. 23. Input mechanical power with the PSS3 and without the PSS3 in the IEEE 9-bus. (a) generator 1, (b) generator 2, (c) generator 3 and (d) generator 4.
Fig. 16 depicts the overall PSO method and how it interplays

with the simulation model during optimization. Fig. 17 depicts
convergence of optimization of algorithms. The optimized
parameters are presented in Table 4. The optimized parameters have
been earned when the mechanical power of G3 has been changed
20% for a four-cycle disturbance, and with considering the variation
of the magnitude and phase of the voltage of the bus 3 (Fig. 15) and
an operating condition (P3 = 0.85 pu, Q3 = 0.11 pu-Table 1).
DSMIB model. Figs. 17 and 18 depict convergence of optimization

of algorithms. The optimized parameters are presented in Table 4.
The optimized parameters have been earned when the mechanical
power of G3 has been changed 20% for a four-cycle disturbance and
normal operating condition (P3 = 0.85 pu, Q3 = 0.11 pu-Table 1).
5.3.2. Designing the PSS3 based on SMIB model (regular model)

After determining the parameters of the PSS3 by the proposed
model (DSMIB model), in this section for understanding the performance of the DSMIB model, the parameters of the PSS3 are designed by the SMIB model. All conditions are as the same of the
6.1. IEEE 9-bus
6. Simulation and results
6.1.1. Linear simulation

In the before section, the parameters of the PSS3 have been designed by two models which are the DSMIB model (proposed
218
model) and SMIB model (regular model), this section shows differences between the performance of these models in enhancing the
power system stability. The dynamic response of the rotor angle
x 10
deviation, angular velocity deviation, internal voltage deviation

and terminal voltage deviation are shown in Fig. 19 (Just the
mechanical power of G3 has been changed 20% for a four-cycle
disturbance for two models) and Fig. 20 (in addition to changing
the mechanical power of G3, DV3, and Dh3 (Table 3) have been
applied to the models). It is clearly seen that the PSS3 which has
been designed by the DSMIB model for all states improved and
enhanced the electromechanical damping characteristics of G3.
-3
PBus7 Bus8,SMIB-With PSS3 regular design

PBus7 Bus8-DSMIB-With PSS3 proposed design
-7
6.1.2. Nonlinear simulation

After obtaining the parameters of the PSS3 by linear models
(both DSMIB and SMIB models), it is essential to investigate the
performance of the proposed design and approach in a nonlinear
power system. In this occasion, IEEE 9-bus has been used. The
low-frequency oscillation in a bulk power system is related to inter-area modes. However, it is difcult to demonstrate the cause
-8
-9
10
15
20
Time (s)
Fig. 24. Flow power from bus 7 to bus 8 with the PSS3 and without the PSS3 in the
IEEE 9-bus.
0.6
Syn 1,DSMIB
0.5
Load Angle (rad)
Load Angle (rad)
Syn 1,SMIB
0.4
0.3
5
10
15
20
25
Syn 2,DSMIB
-0.15
-0.2
-0.25
-0.3
-0.35
-0.4
30
10
15
Time (s)
(a)
20
0.25
0.2
0.15
0.1
0.05
0
0
30
-0.3
-0.35
-0.4
15
20
25
10
10
15
20
(c)
-0.25
Syn 3,DSMIB
(b)
Syn 4,DSMIB
Time (s)
Syn 4,SMIB
-0.2
-0.45
25
Syn 3,SMIB
0.3
Time (s)
Load Angle (rad)
Load Angle (rad)
-0.15
Syn 2,SMIB
-0.1
Load Angle (rad)
P bus 7 to bus 8 (pu)
PBus7 Bus8-Without PSS3
-6
Syn 5,SMIB
-0.15
30
Syn 5,DSMIB
-0.2
-0.25
-0.3
-0.35
-0.4
-0.45
0
30
25
10
15
Time (s)
20
25
30
Time (s)
(d)
(e)
1.0004
1.0002
1
0.9998
0.9996
10
15
20
1.0004
1.0002
1
0.9998
0.9996
10
15
Syn 3,DSMIB
1.002
Syn 4,SMIB
Syn 4,DSMIB
1.001
1.0005
1
0.9995
5
10
Time (s)
(d)
15
1.0002
1
0.9998
0.9996
20
10
15
20
25
Time (s)
(b)
1.0015
Syn 3,SMIB
1.0004
Time (s)
(a)
Syn 2,DSMIB
1.0006
Time (s)
0.999
1.0006
Syn 2,SMIB
1.0008
Angular Velocit (pu)
Syn 1,DSMIB
Syn 1,SMIB
(c)
1.0006
Fig. 25. Load angle with the PSS4 in the IEEE 14-bus. (a) generator 1, (b) generator 2, (c) generator 3, (d) generator 4, and (f) generator 5.
Syn 5,SMIB
1.001

Syn 5,DSMIB
1.0005
1
0.9995
0
10
15
20
25
30
Time (s)
(e)
Fig. 26. Angular velocity deviation with the PSS4 in the IEEE 14-bus. (a) generator 1, (b) generator 2, (c) generator 3, (d) generator 4, and (f) generator 5.
30
of the happening of the low-frequency oscillations in the bulk

power system. In the power system model, inter-area and local
low-frequency oscillations have been observed by applying a fault
at 0 s for 4/60 s (4 cycles) at bus 8 in the IEEE 9-bus standard system (Fig. 1).
Figs. 2123 clearly show that installation and designing PSS3
based on the proposed approach (allocation of PSSs) and linear
model (DSMIB), is a very effective way to better damp lowfrequency oscillations and lead to enhance electromechanical
damping characteristics of the system in comparison with another
case in this study. The power ow between bus 7 and bus 8 which
plays an important role on the network interconnection is shown
in Fig. 24. It can be seen that the installation of the PSS3 can be
enhance the system dynamic performance well and lines can
transmit the power stably.
6.2. IEEE 14-bus
For validating the performance of the proposed approach (allocation of PSSs) and linear model (DSMIB), another nonlinear simulation has been performed. As shown in Fig. 2, IEEE 14-bus
standard system has been used for this reason. The proposed approach has been utilized for determining the dominant generator
in this power system (as a result G4 has been chosen for supplying
PSS) and the proposed linear model (DSMIB) has been used for
determining the parameters of the PSS4. In this power system
model, inter-area and local low-frequency oscillations have been
observed by breaking the line 24 at 0 s for 4/60 s (4 cycles)
(Fig. 1).
Figs. 25 and 26 clearly depict that installation and designing PSS
based on the proposed approach and linear power system model, is
a very effective way to better damp low-frequency oscillations and
lead to enhance electromechanical damping traits of the system. As
a result, it can be concluded that the proposed method can apply to
any large power systems.
START
Randomly Population Initialization

(Position X(0) and Velocity V(0))
Fitness Value
Evaluation
i=i+1
If fitness (X(i))>Fitness (pbest)

Pbest(i)=X(i)
Update Position
By Equation (A1)
If fitness (X(i))>Fitness (gbest)

gbest(i)=X(i)
Update Velocity
By Equation (A2)
i<N
The best particle is selected
END
Fig. A1. PSO ow chart.
219
7. Conclusion
This paper has been proposed an approach and a new linear
model (DSMIB) to allocate and design power system stabilizers
with regard to the local and inter-area modes in order to damp
power system oscillations. Allocation and design procedures of
PSSs have been applied to the IEEE 9-bus standard test system.
Based on this, the eigenvalues analysis method has been introduced to identify dominant generators for installing PSSs. Among
the eigenvalues identied, two dominant modes, i.e. low-frequency dominant mode and stability dominant mode have been
focused and stabilized. The proposed approach succeeded in determining the dominant generator with the ability to inuence these
modes by evaluating the participation factors to the low-frequency
dominant and stability dominant modes. After nding the dominant generator, the PSS should be designed based on the linear
model of this generator, therefore this paper proposed a new linear
model (DSMIB model) including oscillation modes of the power
system. Simulation results show that the PSS3 based controller
which has been designed by the DSMIB model achieves good robust performance, provides superior damping in comparison with
the PSS3 based controller which has been designed by the SMIB
model and enhance greatly the dynamic stability of power systems. Moreover, for validating the proposed method, this procedure has been applied to a larger power system the IEEE 14-bus
test standard system. The obtained results in this test system show
that the system composed with the proposed controller can provide an excellent capability in fast damping of power system oscillations and improve greatly the dynamic stability of power system.
Finally, it can be said that the damping of the SMIB power system can be signicantly enhanced with DSMIB model and the proposed procedure can be applied to any power systems.
Appendix A. PSO algorithm
PSO is a population based stochastic optimization technique
developed by Eberhart and Kennedy in 1995 [24]. The PSO algorithm is inspired by social behavior of bird ocking or sh schooling. The standard PSO algorithm employs a population of particles.
The particles y through the n-dimensional domain space of the
function to be optimized (in this paper, minimization is assumed).
The state of each particle is represented by its position xi = (xi1, xi2, . . . , xin) and velocity vi = (vi1, vi2, . . . , vin), the states of the particles
are updated. The ow chart of the procedure is shown in Fig. A1.
During every iteration, each particle is updated by following two
best values. The rst one is the position vector of the best tness.
This particle has achieved so far. The tness value pi = (pi1, pi2, . . . , pin) is also stored. This position is called pbest. Another best
position that is tracked by the particle swarm optimizer is the best
position, obtained so far, by any particle in the population. This
best position is the current global best pg = (pg1, pg2, . . . , pgn) and
is called gbest. At each time step, after nding the two best values,
the particle updates its velocity and position according to Eqs. (A1)
and (A2), respectively.
v ik1 x v ik c1 r1 pbesti xik c2 r2 gbestk xik

xik1 xik v ik1
A1
A2
where vik+1 is the velocity of particle number (i) at the (k + 1)th iteration, xik is the current particle (solution or position). r1 and r2 are
random numbers between 0 and 1. c1 is the self condence (cognitive) factor; c2 is the swarm condence (social) factor. Usually c1
and c2 are in the range from 1.5 to 2.5; x is the inertia factor that
takes values downward from 1 to 0 according to the iteration number. When a predetermined termination condition is reached, pg is
returned as the optimal value found.
220
References
[1] Dysko A, Leithead WE, OReilly J. Enhanced power system stability by
coordinated PSS design. IEEE Trans Power Syst 2010;25:41322.
[2] Johansson N, ngquist L, Nee H. An adaptive controller for power system
stability improvement and power ow control by means of a thyristor
switched series capacitor (TSSC). IEEE Trans Power Syst 2010;25:38191.
[3] Ayres HM, Kopcak I, Castro MS, Milano F, da Costa VF. A didcate procedure for
designing power oscillation dampers of FACTS devices. Simul Model Practice
Theory 2010;18:896909.
[4] Khodabakhshian A, Hemmati R. Robust decentralized multi-machine power
system stabilizer design using quantitative feedback theory. Electric Power
Energy Syst 2012;41:1129.
[5] Radaideh SM, Nejdawi IM, Mushtaha MH. Design of power system stabilizers
using two level fuzzy and adaptive neuro-fuzzy inference systems. Electric
Power Energy Syst 2012;35:4756.
[6] Hassan LH, Moghavvemi M, Almurib HAF, Muttaqi KM, Du H. Damping of lowfrequency oscillations and improving power system stability via auto-tuned PI
stabilizer using TakagiSugeno fuzzy logic. Electric Power Energy Syst
2012;58:7285.
[7] Hugan X, Haozhong Ch, Haiyu L. Optimal reactive power ow incorporating
static voltage stability based on multi-objective adaptive immune algorithm.
Energy Convers Manage 2008;49:117581.
[8] Kunder P. Power system stability and control. New York, USA: McGraw-Hill;
1994. 0-07-035958-X.
[9] Anderson PM, Fouad AA. Power system control and stability. second
ed. Manhattan, USA: John Wiley & Sons; 2003. 978-0-471-23862-1.
[10] Bikash P, Balarko C. Robust control in power system. New York, USA: Springer
Science & Business Media; 2005. 0-387-25949-X.
[11] Schleif FR, White JH. Damping for the northwestsouthwest tie-line
oscillations an analogue study. IEEE Trans Power Appar Syst 1966;85:
123947.
[12] DeMello FP, Concordia C. Concepts of synchronous machine stability as affected

by excitation control. IEEE Trans Power Appar Syst 1969;88:31629.
[13] Tse GT, Tso SK. Renement of conventional PSS design in multi-machine
system by modal analysis. IEEE Trans Power Syst 1993;8:598605.
[14] Kundur P, Klein M, Rogers GJ, Zywno MS. Application of power system
stabilizers for enhancement of overall system stability. IEEE Trans Power Syst
1989;4:61426.
[15] Ishimaru M, Yokoyama R, Neto OM, Lee KY. Allocation and design of power
system stabilizers for mitigating low-frequency oscillations in the eastern
interconnected power system in Japan. Electrical Power Energy Syst
2004;26:60718.
[16] http://psdyn.ece.wisc.edu/IEEE_benchmarks.
[17] Mithulananthan N, Caizares CA, Reeve J. Indices to detect Hopf bifurcation in
power systems. In: Proceedings of NAPS-2000; 2000. p. 1523.
[18] . IEEE Trans Power Syst 1993;8:2318.
[19] Khalde M, Sarkar A, Lee K, Park Y. The modal performance measure for
parameter optimization of power system stabilizers. IEEE Trans Energy
Convers 1993;8:6606.
[20] Zhang XP, Rehtanz C, Pal B. Flexible AC transmission systems: modeling and
control. Germany: Springer; 2006. 978-3540306061.
[21] Eszterglayos J, Yee MT, Chamia M, Liberman S. The development and operation
of an ultra high voltage speed relaying system for EHV lines. CIGRE 1978:
434.
[22] Bergen R. Power system analysis. Englwood Cliffs: Prentice-Hall; 1986.
[23] Diao Yi-min, Wasynczuk O, Krause PC. Solution of state equations in terms of
separated modes with applications to synchronous machines. IEEE Trans EC1986;1:2116.
[24] Kennedy J, Eberhart RC. Particle swarm optimization. IEEE Int Conf Neural
Networks 1995;4:19428. Perth, Australia.

Improvement of Low Frequency Oscillation Damping by Allocation and

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Improvement of Low Frequency Oscillation Damping by Allocation and

Загружено:

Авторское право:

Доступные форматы

Electrical Power and Energy Systems 52 (2013) 207220

Contents lists available at SciVerse ScienceDirect

Electrical Power and Energy Systems

Improvement of low frequency oscillation damping by allocation and

Department of Electrical Engineering, Amirkabir University of Technology, Tehran 15916-34311, Iran

specic power plant has an inuence on the system. This type of

analysis is efcient in appraising the system stability for a

where x is the vector of state variables, such as rotor angle and

Dx_ ADx BDu;

matrix that denes the proportion of input which appears directly

One problem in using right and left eigenvectors individually for

Fig. 1. IEEE 9-bus test system.

e0di xqi  x0qi 

e00qi e0qi  x0di  x00di

xqi  x0qi iqi

xqi  x0qi iqi

where the electrical power output of the ith generator is as follows:

The governor and AVR of the generators G1 and G4 are dened as

mode oscillations. Local problems may also be associated with

4. Allocation of power system stabilizers

4.2. Eigenvalue analysis of the rst case study

Fig. 2. IEEE 14-bus test system.

Fig. 4. System dominant eigenvalues in IEEE 9-bus test system.

Fig. 5. Linearized model of the SMIB system [15].

4.3. Determination of the dominant generators for installing PSSs by

Mode 6) is selected for the main dominant inter-area mode. In this

Fig. 9. The Single-machine-innite-bus system.

5. Designing PSS based on linear model with focus on the

5.1. Power system models for studying low-frequency oscillations

Fig. 11. The Developed-Single-machine-normal-bus system

The SMIB system is a simple model but in this model, inter-area

Fig. 12. Phase diagram of the ith generator.

already provided with generators. PSSs are designed by linearized

5.2. Developed Single Machine Innite Bus Systems (DSMIB)

Fig. 13. Block diagram of the DSMIB system.

0.1sin(4t)  0.01sin(6t) + 10sin(10t)

Inter-area and local modes

Update the Best

Check the system

Fig. 16. Optimization method on stochastic simulation.

Fig. 18. Convergence of PSO for the SMIB model.

The above equations can be linearized for small oscillation

Pgi V j E0qi =x0di sindi  hj 1=xqi  1=x0di V 2j =2 sin 2di  hj

Fig. 17. Convergence of PSO for the DSMIB model.

(like the multi-machine power system) or the impacts of the rest of

V i V j jxe Iqi  jIdi ejdi

Iqi V j Sindi  hj =xqi

Moreover if Eqs. (22)(24) are linearized around a certain operating

DPgi k1 Ddi k2 DE0qi k1 Dhj k2 DV j

Angular Velocity Deviation (pu)

Load Angle Deviation (rad)

Terminal Voltage Devation (pu)

Internal Voltage Deviation (pu)

Therefore, if Eqs. (22)(24) are substituted into Eqs. (18)(21) and

5.3. Designing PSS

E0qi =x0di sindi

K 4 xdi  x0di =x0di V j sindi  hj

x0di =x0di cosdi

In this section, after proposing the DSMIB model for designing

Terminal Voltage Deviation (pu)

Internal Voltage Deviation (pu)

Angular Velocity Deviation (pu)

Load Angle Deviation (rad)

e0di xqi x0qi

e00qi e0qi x0di x00di

xqi x0qi iqi

xqi x0qi iqi

0.1sin(4t) 0.01sin(6t) + 10sin(10t)

Pgi V j E0qi =x0di sindi hj 1=xqi 1=x0di V 2j =2 sin 2di hj

V i V j jxe Iqi jIdi ejdi

Iqi V j Sindi hj =xqi

E0qi =x0di sindi

K 4 xdi x0di =x0di V j sindi hj

x0di =x0di cosdi