Optimization of Power System Stabilizers Relying On Particle Swarm Optimizers

Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

ICGST-ACSE Journal, ISSN 1687-4811, Volume 8, Issue II, December 2008

Optimization of Power System Stabilizers Relying on Particle Swarm


Optimizers
* A. Jeevanandham ** K.Thanushkodi
* Research Scholar, Anna University, Chennai, India
jeeva932@yahoo.com
** Principal, Coimbatore Institute of Engineering and Information Technology, Coimbatore, India
thanush_dr@rediffmail.com
phase compensation and introduced as a lead-lag
compensator [2].
An inter connected power system, depending on its size,
has hundreds to thousands of modes of oscillation. In the
analysis and control of system stability, two distinct types
of system oscillations are usually recognized. One type is
associated with units at a generating station swinging
with respect to the rest of the power system. Such
oscillations are referred to as local plant mode
oscillations. The frequencies of these oscillations are
typically in the range of 0.8 to 2.0 Hz. The second type
of oscillations is associated with the swinging of many
machines in one part of the system against machines in
the other parts. These are referred to as inter area mode
oscillations, and have frequencies in the range of 0.1 to
0.7 Hz. The basic function of PSS is to add damping to
both types of system oscillations. Other modes which
may be influenced by PSS include torsional modes and
control modes such as the exciter mode associated with
the excitation system and the field circuit [3]. The over
all excitation control system is designed so as to
maximize the damping of the local plant mode as well as
inter-area mode oscillations with out compromising the
stability of other modes and to enhance system transient
stability.
Input to PSS is rotor speed deviation which results
damping torque. In this paper the state space model of the
system with PSS and AVR is designed. The time
constants in the PSS are tuned using simulated annealing
and particle swarm optimization. From the simulation
results, the optimum design of PSS is obtained by using
the best technique.

Abstract
A classical lead-lag power system stabilizer is used for
demonstration in this paper. Initially single first-order
phase compensation block is considered. The stabilizer
parameters are selected in such a manner to damp the
rotor oscillations. The problem of selecting the stabilizer
parameters is converted to a simple optimization problem
with an eigen value based objective function and it is
proposed to employ simulated annealing and particle
swarm optimization for solving optimization problem.
The objective function allows the selection of the
stabilizer parameters to optimally place the closed-loop
eigen values in the left hand side of the complex s-plane.
The effectiveness of the stabilizer tuned using the best
technique, in enhancing the stability of power system.
The performance of the system is analyzed by using both
the techniques. Stability is confirmed through eigen
value analysis and simulation results and suitable
heuristic technique will be selected for the best
performance of the system. 1
Keywords: Rotor oscillations, Power system stability,
Robust control, Simulated Annealing, Particle swarm
optimization.

1. Introduction
During changes in operating conditions, oscillations of
small magnitude and low frequency often persist for long
period of time and in some cases even present limitations
on power transfer capability. Power system stabilizer
(PSS) is designed to damp the low frequency oscillations
of power system [1].
PSS is used to add damping to the generator rotor
oscillations by controlling its excitation using auxiliary
stabilizing signals. The widely used conventional power
system stabilizer (CPSS) is designed using the theory of

2. Controller Design
Damping torque is produced to overcome rotor
oscillation. The action of a PSS is to extend the angular
stability limits of a power system by providing
supplemental damping to the oscillation of synchronous
machine rotors through the generator excitation [4].

This study has been implemented on MatLab 7.0


platform at Power System Simulation Lab, Anna
University (GCT Campus)

ICGST-ACSE Journal, ISSN 1687-4811, Volume 8, Issue II, December 2008

Controller is designed to compensate lag between exciter


input and electrical torque.

optimization problem solved with simulated annealing


and particle swarm optimization.
Consider the problem of determining the parameters of a
stabilizer that relatively stabilizes a family of N plants,

TPSS
= K
VS

X (t ) = Ak X (t ) + Bk U (t ) ; k = 1,2,3,... N (1)
.
Where, X(t) Rn is the state vector and X(t) Rm is
the control vector.
Very often, the closed-loop modes are specified to have
some degree of relative stability. In this case closed-loop
eigen values are constrained to lie to the left of a vertical
line corresponding to a specified damping factor. A
necessary and sufficient condition for the set of plants in
equation (1) to be simultaneously relatively stabilizable
with a single control law is that the eigen values of the
closed-loop system lie in the left-hand side of a vertical
line in the complex s-plane. This condition motivates the
following approach for determining the parameters of the
PSS.
Select the parameters of the PSS to minimize the
following objective function:

The amount of damping introduced depends on the gain


of PSS transfer function at that particular frequency of
oscillation.
r

VS

K STAB

Gain

sTW
1 + sTW
Washout

1 + sT1
1 + sT2
Phase Compensation

Figure 1. Lead Lag Power System Stabilizer

Phase Lead Compensation:


To provide damping, the stabilizer must produce a
component of electrical torque which is in phase with
speed variations. Therefore, the PSS transfer function
should have an appropriate phase-lead characteristics to
compensate for the phase lag between the exciter input
and the electrical torque.
The phase characteristic to be compensated, changes with
system conditions. Therefore, compromise must be made
and a characteristic acceptable for a desired range of
frequencies (normally 0.1 to 2.0 Hz) and for different
system conditions is selected. This may result in less than
optimum damping at any one frequency. Generally, slight
under compensation is preferable to overcompensation so
that both damping and synchronizing torque components
are increased.
Stabilizing Signal Washout:
The signal washout function is a high pass filter which
removes dc signals, and without it steady changes in
speed would modify the terminal voltage. The washout
time constant is in the range of 1 to 20 seconds. For local
mode oscillation, a wash out of 1 to 2 sec is satisfactory.
From the view point of low frequency inter area
oscillations a washout time constant of 10 sec or higher
may be required in order to reduce phase lead at low
frequencies.
Stabilizer Gain:
The stabilizer gain (KSTAB) is chosen by examining the
effect for a wide range of values. Ideally, the stabilizer
gain should be set at a value corresponding to maximum
damping. Gain is set to a value which results in
satisfactory damping of the critical system mode(s)
without compromising the stability of other modes, or
transient stability, and which does not cause excessive
amplification of stabilizer input signal noise.
Stabilizer Output Limits:
In order to restrict the level of generator terminal voltage
fluctuation during transient conditions, limits are
imposed on the PSS output. The effect of the two limits is
to allow maximum forcing capability while maintaining
the terminal voltage within the desired limits [4].

J = max{Re( k ,i ) + } ; k = 1,2,3,... N ; i = 1,2,...n

where k,i is ith closed loop eigen value of the kth plant
and is relative stability factor. Subject to the constraints
that finite bounds are placed on the stabilizer parameters.
In this paper instead of N number of plants, singlemachine-infinite-bus system is considered.
The objective function can be modified as,

J = max{Re( k ) + }

The relative stability is determined by the value of as


shown in figure.2

Figure 2. Region in the left-hand side of a vertical

If any solution is found such that J < 0, then the resulting


stabilizer parameters relatively stabilize the oscillations
in the collection of plants. The existence of a solution is
verified numerically [5] by minimizing J.

4. System Model
In this steady single-machine-infinite-bus power system
is considered [6]. The supplementary stabilizing signal
considered is one proportional to speed [7].A widely used
conventional PSS is considered throughout the study.
The transfer function of PSS with single phase
compensation block is

3. Problem Formulation
In this section, the eigen value-based objective function
used to robustly select the PSS parameters [5], and the
8

ICGST-ACSE Journal, ISSN 1687-4811, Volume 8, Issue II, December 2008

VS
sTW 1 + sT1
= K STAB
r
1 + sTW 1 + sT2

A parameter Cp is introduced to control the temperature T.


The basic elements of SA are
Trial, Current and Best Solutions:
xtrial, xcurrent, and xbest are set of the optimized parameter
values at any iteration
Acceptance Criterion:
At any iteration, the trial solution can be accepted as the
current solution if it meets one of the following criteria
(a) J ( xtrial ) p J (xcurrent )
(b) J (x trial ) f J (x current )

The first term is stabilizer gain. The second term is


washout term with a time lag Tw. The third term is a lead
compensation [4] to improve the phase lag through the
system.
The numerical values of Tw, T2 and system data are
given in Appendix I. The remaining parameters namely
KSTAB and T1 are assumed to be adjustable parameters.
The optimization problem is selection of these PSS
parameters easily and accurately. The optimization
problem can be solved using the simulated annealing as
well as particle swarm optimization. The SA algorithm in
explained in Appendix II and PSO algorithm is explained
in Appendix III.
For a given operating point, the power system is
linearized around the operating point, the eigen values of
the closed-loop system are computed, and the objective
function is evaluated. It is worth mentioning that only the
system electromechanical modes are incorporated in the
objective function. The bounds on the parameters used in
the SA are given in Appendix I.

PSS

K4

Ga

Ge

Efd
+

Tm

G fd

Gf

K2
K6

377
- Te +

Gm
r S
+

Fig. 4 System Model

K1

+
+

)) / C p ) rand (0,1)

Here J(xtrial) and J(xcurrent) are the values of objective


function corresponding to xtrial and xcurrent respectively.
And rand (0,1) is a set of random number with
domain (0,1).
Acceptance Ratio:
At a given value of Cp, an n1 trial solutions can be
randomly generated. Based on the acceptance criterion,
an n2 of these solutions can be accepted. The acceptance
ratio is defined as n2/n1.
Cooling Schedule:
It specifies a set of parameters that governs the
convergence of the algorithm. This set includes an initial
value of control parameter Cp0, a decrement function for
decreasing the value of Cp, and a finite number of
interactions or transitions at each value of Cp i.e. the
length of each homogeneous Markov Chain. The initial
value of Cp should be large enough to allow virtually all
transitions to be accepted. However, this can be achieved
by starting off at a small value of Cp0 and multiplying it
with a constant larger than 1, i.e. Cp0 = Cp0. This
process continues until the acceptance ratio is close to 1.
This is equivalent to heating up process in physical
systems. The decrement function for decreasing the value
of Cp is given by Cp = Cp where is a constant close
to 1. Typical values lie between 0.8 0.9 [8].
Stopping Criteria:
These are the conditions under which the search process
will terminate. In this study the search will terminate if
any one of the following criteria is satisfied,
(a) The number of Markov Chains since the last change
of the best solution is greater than a prespecified number
(b) The number of Markov Chains reaches the maximum
allowable number.

VS
Vref +

(c) exp ( J ( x trial ) J ( x current

K5

Figure 3. Block diagram of AVR with PSS

5. Simulated Annealing
Simulated Annealing [8] is an optimization technique
that simulates the physical annealing process in the field
of combinatorial optimization. Annealing is the physical
process which involves heating up of a solid until it melts,
followed by slow cooling down by decreasing
temperature. Thermal equilibrium at any temperature T
i.e temperature of solid is maintained as constant for a
period of time, is given by Boltzmann distribution. It
gives the probability of the solid being in an energy state
i with energy Ei at temperature T as,

6. Particle Swarm Optimization


Similar to evolutionary algorithms, the PSO technique
conducts searches using a population of particles,
corresponding to individuals. In a PSO system [9],
particles change their positions by flying around in a
multidimensional search space until a relatively changed
position has been encountered, or until computational
limitations are exceeded.
The following are the advantages of PSO [10] over other
traditional optimization techniques like classical cost
function and genetic algorithms,
a) PSO is a population based search algorithm.
This property ensures it to be less susceptible to
getting trapped on local minima.

Ei

Pi = k e T
where k is Boltzmanns constant. The analogy between a
physical annealing process and a combinatorial
optimization problem is based on the following,
Solutions obtained for optimization problem are
equivalent to configurations of a physical
system
The cost of a solution is equivalent to the energy
of a configuration.
9

ICGST-ACSE Journal, ISSN 1687-4811, Volume 8, Issue II, December 2008

b) PSO uses payoff (performance index or


objective function) information to guide the
search in the problem space.
c) PSO uses probabilistic transition rules and not
deterministic rules. Hence, PSO is a kind of
stochastic optimization algorithm that can
search a complicated and uncertain area.
Particle X(t):
It is a candidate solution presented by an m-dimensional
real-valued vector, where m is the number of optimized
parameters.
Population pop(t):
It is a set of n particles at time t. (i.e. pop(t) = [X1(t),
X2(t), , Xn(t)]T
Swarm:
It is disorganized population of moving particles that tend
to cluster together while each particle seems to be
moving in a random direction.
Particle velocity V(t):
It is the velocity of the moving particles represented by
an m-dimensional real valued vector. At the time t, the jth
particle velocity Vj(t) = [vj,1(t), vj,2(t),., vj,m(t)], where
vj,k(t) is the velocity component of jth particle with
respect to the kth dimension.
Individual best X*(t):
As the particle moves through the search space, it
compares its fitness value at the current position to the
best fitness value it has ever attained at any time up to the
current time. The best position that is associated with the
best fitness encountered so far is called is called
individual best X*(t). For each particle in the swarm, X*(t)
can be determined and updated during the search. In a
minimization problem objective function J, the individual
best of the jth particle Xj*(t) is determined so that J(Xj*(t))
J(Xj*(), t. For simplicity, assume that Jj* = J(Xj*(t)).
For the jth particle, individual best can be expressed as
Xj(t) = [x*j,1(t), x*j,2(t),., x*j,m(t)].
Global best X**(t):
It is the best position among all of the individual best
positions achieved so far. Hence, the global best can be
determined such that J(X**(t)) J(Xj*(t)), j = 1,2,n. For
simplicity, assume that J** = J(X**(t)).
Stopping criteria:
These are the conditions under which the search process
will terminate. In this study, the search will terminate if
one of the following criteria is satisfied,
a) The number of iterations since the last change of the
best solution is greater than a prespecified number
b) The number of iterations reaches the maximum
allowable number.
The particle velocity in the kth dimension is limited by
some maximum value, vkmax. This limit enhances the
local exploration of the problem space and it realistically
simulates the incremental changes in human learning.
The maximum velocity in the kth dimension is
characterized by the range of the kth optimized parameter
and given by
x max x kmin
v kmax = k
N
where N is a chosen number of intervals in the kth
dimension.

7. Simulation Results
In this part of the study, a single machine is connected to
infinite bus through a transmission line, and operating at
different loading conditions [11], is considered. The
linearized model of this system, voltage regulator and
exciter included [6] is considered.
The constants K1 to K6 , with the exception of K3 , which
is only a function of the ratio of impedance, are
dependent on the actual real power (P) and reactive
power (Q) loading as well as the excitation levels in the
machine.
The operating points are selected based on the different
loading conditions. The simultaneous damping
enhancement of the is demonstrated by considering five
different loading conditions.
The operating points were selected randomly as follows:
( Po , Qo ) = (0.9, 0.3); (0.8, -0.1); (0.5,0.5); (0.6, -0.2);
(1.0, 0.6)
The eigen values are found by transferring the transfer
function of the system data into state model.
The eigen values of the system at the five operating
points considered, with out PSS are,
1. 0.4981 6.6288i, -33.6805, -17.3597
2. 0.7513 7.3702i, -11.5526, -39.9942
3. 0.0283 5.3580i, -25.0504 9.1822i
4. 0.1936 6.9157i, -10.7786, -39.6528
5. 0.5410 6.1171i, -21.6341, -29.4919
From the eigen values it is clear that the system is
unstable due to its location in the right half of the
s plane.
Plot for System w ithout PSS
0.25
0.2

Load angle amplitude

0.15
0.1
0.05
0
-0.05
-0.1
-0.15
-0.2

0.5

1.5

2.5

3.5

4.5

Time (sec)

Figure 4. Response of the system without PSS

For the system to be stable the real part of eigen value


should be located in the in the left hand side of s-plane.
Results for Simulated Annealing:
Simulated annealing is used to optimize the objective
function J for five different kind of operating conditions
to shift the eigen values to the left vertical line.
The eigen values of the five systems, with PSS are
1. -46.3181,-16.829543.2325i,-0.1883 2.8091i,-0.7077
2. -45.7209,-16.475528.1318i,-0.8262 3.8261i,-0.7372
3. -35.6015,-21.698613.9605i,-0.6626 4.9864i,-0.7376
4. -44.0326,-16.532918.3720i,-1.6023 4.3745i,-0.7585
5. -45.6105,-17.218340.4834i,-0.1521 2.8783i,-0.7102

10

ICGST-ACSE Journal, ISSN 1687-4811, Volume 8, Issue II, December 2008

System response w ith PSS for P = 1.0 and Q = 0.6


0.05

0.04

0.04

0.03

0.03

0.02

Load angle amplitude

Load angle amplitude

System response w ith PSS for P = 0.9 and Q = 0.3


0.05

0.01
0
-0.01

0.02
0.01
0
-0.01
-0.02

-0.02
-0.03

-0.03
-0.04

-0.04

10

15

20

25

10

15

20

30

25

30

35

40

Time (sec)

Time (sec)

Figure 5. Response of the system with PSS using


Simulated Annealing

System response w ith PSS for P = 0.8 and Q = -0.1


0.035

From figure.5 it is obvious that the rotor oscillations are


damped and the above results shows that all the eigen
values of the system were located in the left half of the
s plane to make the system stable.

0.03
0.025

Load angle amplitude

0.02
0.015
0.01

Results for Particle Swarm Optimization:


The objective function J is optimized with the PSO and
N=5 to shift the electromechanical mode of each of the
five systems to the left of the vertical line defined by
= -2.2.
The eigen values of the five systems, with PSS are
1. -34.16705.0796i, -9.9855, -0.9887 7.7731i, -0.7644
2. -37.4972,-35.6751,-0.5533 9.3491i, -6.0033, -0.7792
3. -29.0306 7.0224i,-20.4992,-0.8755 5.6109i,-0.7500
4. -36.4206 0.9044i,-1.3020 9.2146i, -4.8165, -0.7998
5. -33.0536 5.8900i,-12.5555,-0.8183 6.9006i, -0.7620
All the eigen values of the system were located in the left
half of the s plane. So the system will be stable.

0.005
0
-0.005
-0.01
-0.015

Time (sec)

System response w ith PSS for P = 0.5 and Q = 0.5


0.03
0.025

Load angle amplitude

0.02
0.015
0.01
0.005
0

System response w ith PSS for P = 0.9 and Q = 0.3


0.025

-0.005
-0.01

0.02
-0.015

Load angle amplitude

Time (sec)

System response w ith PSS for P = 0.6 and Q = -0.2


0.03

0.025

0.015

0.01

0.005

0
0.015

-0.005
0.01

Time (sec)

0.005

System response w ith PSS for P = 0.8 and Q = -0.1


0.025

-0.005

0.02

Time (sec)

Load angle amplitude

Load angle amplitude

0.02

0.015

0.01

0.005

-0.005

4
Time (sec)

11

ICGST-ACSE Journal, ISSN 1687-4811, Volume 8, Issue II, December 2008

Table 1: Comparison between simulated annealing


and Particle Swarm Optimization

System response w ith PSS for P = 0.5 and Q = 0.5


0.03
0.025

Loading
Condition
( Po , Qo )

Load angle amplitude

0.02
0.015

Simulated
Annealing
Settling
Time
in Sec.

Peak
Amp.

Particle Swarm
Optimization
Settling
Time
in Sec.

Peak
Amp.

0.01

0.9, 0.3

20.8

0.0443

4.51

0.0236

0.8, -0.1

4.6

0.0313

6.32

0.0211

0.5, 0.5

5.58

0.0287

4.92

0.0275

0.6, -0.2

3.96

0.0272

5.06

0.0218

1.0, 0.6

4.09

0.0324

4.99

0.026

0.005
0
-0.005

-0.01

Time (sec)

System response w ith PSS for P = 0.6 and Q = -0.2


0.025

Load angle amplitude

0.02

From the above comparison the simulated annealing is


having higher settling time, higher peak amplitude and
higher computational time than Particle swarm
optimization. So tuning of PSS parameters by using PSO
is more optimal than SA, Classical Cost Function and
Genetic Algorithm [12]. If we considered N plants, the
same results hold good for individual plants.

0.015

0.01

0.005

8. Conclusion

The use of Simulated Annealing and PSO to design


robust power system stabilizers for power systems
working at various operating conditions are investigated
in this paper. The problem of selecting the PSS
parameters, which simultaneously improve the damping
at various operating conditions, is converted to an
optimization problem with an eigen value-based
objective function which is solved by both Simulated
Annealing and PSO. For a general loading condition with
real power 0.9 and reactive power 0.3, the simulated
annealing will settle down oscillations of the system
under study at 20.8 seconds with peak amplitude of
0.0443, whereas PSO will settle down oscillations at 4.51
seconds with peak amplitude of 0.0236. So it is evident
that PSO is more optimal than SA for this case.
The objective function is presented allowing the robust
selection of stabilizer parameters that will optimally
place the closed-loop eigen values in the left hand side of
a vertical line in complex s-plane. By comparing the
above two meta-heuristic optimization techniques, it is
found that Particle Swarm Optimization is better than
simulated annealing in tuning the parameters of the
Power system stabilizer, to reduce intra and inter area
rotor oscillations over a wide range of operating
conditions.

Time (sec)

System response w ith PSS for P = 1.0 and Q = 0.6


0.03

0.025

Load angle amplitude

0.02

0.015

0.01

0.005

-0.005

Time (sec)

Figure 6. Response of the system with PSS using


Particle Swarm Optimization

Figure.6 shows response of the system with power


system stabilizer for various operating conditions
whereas their parameters were tuned by using Particle
Swarm Optimization (PSO). It is indicating simultaneous
improvement in the response of the five systems.
The above two techniques can be compared as shown in
Table 1,

12

ICGST-ACSE Journal, ISSN 1687-4811, Volume 8, Issue II, December 2008

APPENDIX III
PARTICLE SWARM OPTIMIZATION

Appendix I
System Data

Step (1): Set the time counter t = 0 and generate random


n particles,{Xj(0), j = 1,2,n}, where Xj(0) =
[xj,1(0), xj,2(0),. xj,m(0)]. xj,k(0) is generated by
randomly selecting a value with uniform
probability over the kth optimized parameter
search space [xkmin, xkmax].
Generate randomly initial velocities of all
particles, {Vj(0), j = 1,2,n}, where Vj(0) =
[vj,1(0), vj,2(0),. vj,m(0)]. vj,k(0) is generated by
randomly selecting a value with uniform
probability over the kth optimized parameter
search space [-vkmax, vkmax].
Each particle in the initial population is
evaluated using the objective function, J. For
each particle, set X*j(0)=Xj(0) and J*j=Jj,
j=1,2,,n. Search for the best value of objective
function Jbest. Set the particle associated with
Jbest as the global beat, X**(0), with an objective
function of J**. Set the initial value of the inertia
weight w(0).
Step (2): Update the time counter t = t+1.
Step (3): Update the inertia weight w(t) = w(t-1).
Step (4): Using the global best and individual best, the jth
particle velocity in the kth dimension is updated
according to the following equation:
v j ,k (t ) = w(t )v j ,k (t 1) + c1 r1 ( x *j , k (t 1) x j ,k (t 1)

Single- Machine- Infinite- Bus System:

1
2HS +KD ;
sK f
Gf =
1 + sTF ;

Gm =

KA
sTA + 1 ;
K3
G fd =
1 + sT ' do K 3
GA =

Ge =

1
sTE + 1 ;

The system data are as follows:


Machine (p.u)

xd = 1.7 ; xd = 0.254 ; xq = 1.64 ; 0 = 120 rad/s ;


Td0 = 5.9 sec ; KD = 0 ; H = 2.37 sec
Transmission Line (p.u)

re = 0.02 ;

xe = 0.4

Exciter and Stabilizer

KA = 400 ; TA = 0.05 sec ; KF = 0.025


TF = 1.0 sec ; KE = -0.17 ; TE = 0.95 sec

+ c 2 r2 ( x *j*,k (t 1) x j ,k (t 1))

TW = 10 sec ; T2 = 0.0227 sec

where c1 and c2 are positive constants and r1 and


r2 are uniformly distributed random numbers in
[0,1].
Step (5): Based on the updated velocities, each particle
changes its position according to the following
equation:
x j , k (t ) = v j , k (t ) + x j , k (t 1)
Step (6): Each particle is evaluated according to the
updated position. Jj < J*j, j=1,2,,n, then update
individual best as X*j(t) = Xj(t) and J*j = Jj, and
go to step 7; else go to step 7.
Step (7): Search for the minimum value Jmin among J*j,
where min is the index of the particle with
minimum objective function value, i.e., min {j;
j=1,2,,n}. If Jmin < J** then update global best
as X** = Xmin(t), and J**=Jmin and go to step 8;
else go to step 8.
Step (8): If one of the stopping criteria is satisfied, then
stop, or else go to step 2.

Bounds for the stabilizer adjustable gain and time


constants are [0.01, 10] and [0.03, 1.0] respectively.
APPENDIX II
SIMULATED ANNEALING ALGORITHM

Step (1): Set the initial value of Cp0 and randomly


generate an initial solution xinitial and calculate
its objective function. Set this solution as the
current solution as well as best solution,
i.e. xinitial = xcurrent = xbest.
Step (2): Randomly generate an n1 of trial solutions in
the neighborhood of the current solution.
Step (3): Check acceptance criterion of these trial
solutions and calculate the acceptance ratio. If
the acceptance ratio is close to 1 go to step 4,
else set Cp0 = Cp0, > 1, and go back to step 2.
Step (4): Set the chain counter kch = 0.
Step (5): Generate a trial solution xtrial. If xtrial satisfies
the acceptance criterion, set xcurrent, = xtrial,
J(xcurrent,) = J(xtrial), and go to step 6;
else go to step 6.
Step (6): Check the equilibrium condition. If it is
satisfied go to step 7, else go to step 5.
Step (7): Check the stopping criteria. If one of them is
satisfied then stop; else set kch = kch +1 and
Cp = Cp, < 1, and go back to step 5.

9. References
[1] M.A. Abido, Robust design of multi machine power
system stabilizers using simulated annealing, IEEE
transactions on Energy Conversion, Vol.15, No.3,
Sep. 2000.
[2] E.V. Larsen and D.A. Swann, Applying Power
System Stabilizers: Parts I, II, and III, IEEE
transactions on Power Apparatus and Systems,
vol.PAS-100, pp. 3017-3046, 1981.

13

ICGST-ACSE Journal, ISSN 1687-4811, Volume 8, Issue II, December 2008

[3] Ali Feliachi, Xiaofan Zhang, Craig S.Sims, Power


System stabilizers design using optimal reduced
order models part I and part II, IEEE transactions
on Power system, vol.3, No.4, Nov 1988.
[4] P. Kundur, M. Klein, G.J. Rogers and M.S. Zywno,
Application of power system stabilizer for
enhancement of over all system stability, IEEE
transaction on Power system, vol.4, No.2, May 1989.
[5] M.A. Abido, Y.L. Abdel-Magid and A.H. Mantawy,
Robust tuning of power system stabilizers in multi
machine power systems, IEEE transaction on Power
systems, vol.15, No.2, May 2000.
[6] P.Kundur, Power system stability and control,
McGraw Hill, 1994
[7] F.P. De Mello and C.Concordia, Concepts of
synchronous machine stability as affected by
excitation control, IEEE transactions on Power
Apparatus and Systems, vol.PAS-92, pp. 1426-4134,
1968.
[8] E. Aarts and J. Korst, Simulated Annealing and
Boltzmaan machines: A stochastic approach to
combinatorial optimization and neural computing
John Wiley and Sons, 1989.
[9] Kennedy, J., Eberhart, R. C., and Shi, Y., Swarm
intelligence San Francisco: Morgan Kaufmann
Publishers, 2001.
[10] Abido, M. A., "Optimal design of power system
stabilizers using particle swarm optimization," IEEE
Transactions on Energy Conversion, vol. 17, no. 3,
pp. 406-413, Sept.2002.
[11] Michhel J. Gibbard, Robust design of fixed
parameter power system stabilizers over a wide
range of operating conditions, IEEE transactions on
Power system, vol.6, No.2, May 1991.
[12] Y.L. Abdel-Magid, M.A. Abido, S. Al-Baiyat and
A.H. Mantawy, Simultaneous stabilization of
multimachine
power
systems
via
genetic
algorithms, IEEE transactions on Power system,
vol.14, pp.1428-1439, Nov 1999.

A. Jeevanandham born in Coimbatore


District, TamilNadu State, India in
1975. He received the B.E. (Electrical
and Electronics Engineering) from
Coimbatore Institute of Technology,
Bharathiar University, Coimbatore,
M.E.(Power Systems Engineering)
from
Government
College
of
Technology, Anna University, Coimbatore and currently
doing Ph.D in the area of Power System Stability at Anna
University, Chennai. His research interests are robust
control of Power system and applications of AI
techniques to Power system.

Dr.K. Thanushkodi, born in Theni


District , TamilNadu State, India in
1948, received the BE in Electrical
and Electronics Engineering from
Madras University, Chennai. MSc
(Engg) from Madras University,
Chennai and PhD in Electrical and
Electronics
Engineering
from
Bharathiar University, Coimbatore in
1972, 1976 and 1991 respectively. His research interests
lie in the area of Computer Modeling and Simulation,
Computer Networking and Power System. He has
published 32 technical papers in National and
International Journals.

14

You might also like