International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 10, October 2013)

Small Signal Stability Analysis of Synchronous Generators

Connected to Infinite Bus
Avijit Maity1, Kesab Bhattacharya2, Amar Nath Sanyal3
1
Asst. Professor, HETC, Hooghly
2
Professor of Electrical Engg., Jadavpur University
3
Professor of Elect. Engg., AOT, Aedconagar (formerly at Jadavpur University
Abstract- A synchronous generator may sometimes be K A ,TA , K F ,TF Constants of exciter model IEEE, type 1s
able to withstand a large impact but it may fail to retain
E ' , Eq ' Voltage behind transient reactance/ its q-axis
synchronism under a small impact due to growing
oscillations, leading to instability. This is called dynamic component
instability. It occurs for specific operating conditions and
network configurations, particularly in presence of fast- I. INTRODUCTION
acting excitation control. The paper presents techniques of
small signal stability assessment of a turbo generator Stability in power system means the ability of
connected to a grid system while a small perturbation is synchronous machines connected to the grid system to
injected externally. The infinite bus idealisation has been retain synchronism between each other. Itmay be
made as the capacity of the generator is small compared to pertaining to either steady state or transient conditions
that of the power grid. The classical model based on Park- and may be either a power angle stability problem or a
Gorev transformation has been used to represent the voltage stability problem. The power angle instability
machines. The current state space model has been used. The arises out of mismatch in active power. The voltage
small signal stability has been assessed by examining the stability problem arises out of mismatch in reactive
eigenvalues of the state matrix. Reduced order model has
also been developed for the same machine including exciter
power [8][9].
parameters and the stability has been examined by applying Steady state stability can again be classified into static
Rouths array on the characteristic equation. and dynamic. While slow changes occur in absence of
control actions, the stability is called static. The dynamic
Index Term-Small signal stability, Current state space, stability problem is, in essence, the performance of a
Linearized model, Reduced order model, Exciter control turbine-generator connected to grid system under small
Symbols used (All quantities are in p.u. and time-constants impacts, in presence of fast acting control systems viz.
in sec. unless otherwise mentioned)- the AVR-excitation control. The small impact may be a
small change in load, or in voltage or in network
, Angular speed of the generator (r/s), power angle(r) parameters[12].
Ra , X a Armature resistance, leakage reactance
A. Modelling Approach
X d , X ds D-axis unsaturated/saturated synchronous reactance
X q , X qs Q-axisunsaturated/ saturated synchronous reactance
In earlier days the turbo generator was modeled as a
voltage source behind its transient reactance. This
X d' , X ds' D-axis unsaturated/ saturated transient reactance approach led to inaccuracy. Now-a-days multivariable
X d" , X ds" D-axis/Q-axis sub-transient reactance systems are modelled in terms of state variables. Either
Re , X e , Ze External resistance,leakage reactance, impedance
current states or flux-linkage states are chosen by the
modeler [7]. The general practice is to use the classical
to infinite bus,
model of Park, Concordia and Adkins for the
Td' , Tdo' D-axis S.C./O.C. transient time constant. synchronous machines [11]. While subjected to small
Td" , Tdo" D-axis S.C./ O.C. subtransient time constant impact the state variables of the system change slightly
"
T ,T "
Q-axis S.C./ O.C. subtransient time constant from the initial statesand assume new states. The initial
q qo
states are called the quiescent states. They can be found
I d , I q D-axis &Q-axis components of armature current out from the known and established performance
Vd ,Vq D-axis & Q-axis components of armature voltage equations of synchronous machine [6][10].
E, EFD Generator Induced e.m.f, exciter output To examine the behaviour of the machine under small
impact, the non-linear state space equations are linearised
Vt ,Vb Voltage at generator terminal, infinite bus
about the quiescent point. The stability is predicted from
K1 K6 Constants of DMello& Concordia the eigenvalues of the linearised state matrix. If real parts
D, H Damping/ Inertia constant of all the eigenvalues are negative, then the system is
stable, otherwise the system is unstable [5][13].

129
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 10, October 2013)
There are two types of non-linearities in the state The load is represented as constant impedance in this
matrix of a machine viz. trigonometric and product treatment.
nonlinearities. They can be eliminated by trigonometric Power station managers are provided with
and algebraic manipulations[7]. manufacturers data from which the equivalent circuit for
The generator-transformer may be treated as one the direct& quadrature axes can be found out. The
machine on infinite bus provided the capacity of the parameters are generally given in p.u. and time-constants
machine is small compared to that of the grid. A in seconds.
multimachine problem can be reduced to a problem of These parameters are with reference to the classical
one machine on infinite bus, by load flow analysis and model of Park, Concordia & Adkins. The effect of
network reduction. However, this approach hampers saturation and distributed eddy currents in the solid iron
accuracy of the solution[1][3]. rotor are not accounted for in the classical model. Such
effects can also be included in the dynamic studies by
II. T HE INITIAL C ONDITIONS extended mathematical models advanced by various
In dynamic studies, the initial conditions of the system authors [11][12].
are required. For one machine on infinite bus with
specified boundary conditions, the pertinent equations for III. LINEARISATION - E IGENVALUE TECHNIQUES F OR
estimation of steady state conditions and the T HE P REDICTIONOF DYNAMIC STABILITY
corresponding phasor diagrams are given in text-books. It has been already stated that for linearisation of non-
The system is shown in fig. 1. The effect of local load linear state space equations [7][13] of transient stability
and transmission line capacitances can be included in the analysis, we have to make trigonometric and algebraic
study [8]. manipulations. If Xo be the initial states & X be the
small departure, then the non-linear state space form:

X = f(X, U, t) (1)

Can be linearised as:

X = A(Xo )X B(Xo )U (2)

The elements of the state-matrix A depends on the

initial states Xo - these are assumed to be constant during
Fig.1 A generator-transformer connected to infinite bus through a
short transmission line the period of interest. The dynamic properties of the
system depend on eigenvalues of the state matrix. Its
For a multimachine system, the bus voltages, active
elements change for a change in quiescent conditions.
and the reactive power injected into the system by a
Anderson and Fouad have derived two matrix forms by
generator or a group of generators are obtained from load
algebraic and trigonometric manipulations. They have
flow study. A group of generators operating in parallel is
used currents as states in the first form and flux-linkages
equivalenced to a single generator for convenience. The
as states in the other. The current state space model is
network is also reduced eliminating unimportant nodes to
given as:
reduce the dimension of the problem. The initial states-
currents, voltages and flux-linkages for individual MX KX - V or X = -M-1K X - M-1 V AX + BU (3)
machines are found out with reference to the slack bus,
by known methods [12]. Expressions for the elements of the M -matrix and the
K -matrix are given in ref[7].
A. Equivalencing Technique And Network Reduction
If the eigenvalues of the state matrix are negative real
While a group of generators swing similarly they form quantities or they are complex with negative real
a coherent group. If there be n-number of generator- components, then the system will remain dynamically
transformers in a coherent group, then their equivalenced stable. Otherwise the system will become unstable.
parameters and time-constants are obtained by well- The alternative state space formulation in flux-
known methods. Equivalencing may also be made by linkages is of the following form:
matching f-response curves.
If we like to retain n- number of nodes & eliminate r- = T-1C + T-1 D AX + BU (4)
number of nodes from the equivalent load admittance
matrix, we have to use network reduction techniques. The expression for the elements of the T and the C
Most often n is the number of equivalenced generators. matrices has also been given by Anderson and Fouad[7].

130
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 10, October 2013)
Both the forms are used for dynamic performance j 2H (12)
studies. These matrices can be extended to include the
effects of: Extensive application of this model has been made in
i. Excitation control various power system problems for its simplicity. A
ii. Turbine-governor control comprehensive report of the applications has been
iii. Power system stabilizer produced by P.M. Anderson et al [7] and P. Kundur [13].
Additional equations are used to include the effects of the
Detailed discussion of such methods has been made in
solid state exciter [7] shown in fig. 2.
the texts [2][12][13]. The dynamic stability is assessed by
finding eigenvalues of the state matrix. The characteristic
equation is found out by Fadeev-Leverrier method and
the eigenvalues by Lin-Bairstow method [14][15].
MATLAB-tools are frequently used for finding out the
eigenvalues and hence to predict the stability of the
system [16]. Its use reduces programming efforts.

E'
IV. S IMPLIFIED LINEAR MODELS - T HE q -M ODEL,
CONCORDIA AND DE MELLO S CONSTANTS Fig. 2. Block diagram representation of exciter- model IEEE 1s.
While some investigators have tried to model a TABLE-I
machine elaborately to get more accurate predictions, [ Type1S model of IEEE has been used, Ref. Fig. 2]
others have tried to reduce the model through idealising
assumptions and approximations[7,12]. Obviously, there Eq' 1 K4 1 x Eq'
is loss of accuracy in prediction with this approach, but K 3Tdo' Tdo' Tdo'
the dimension of the model gets reduced which saves
computer memory space and run time. It is a useful
technique particularly for large interconnected systems. K2 D K1
Individual machines or group of machines remote from j j j
the point of perturbation does not appreciably affect the
system dynamics. Hence it is expedient to use a reduced
model for them. = 1
With this idea in view, a reduced order model was
developed by Concordia and De Mello, based on V1 (T3 K f K a ) K f V1
(E' - E'q ) concept. Their analysis gave rise to a set of T f Ta
TaT f
equations in terms of constants K1 to K 6 . The
expressions for constants of Concordia and De Mello are
given below: EFD Ka 1 EFD
Ta Ta
K1 KiVb [ Eqao {Re sin ( o ) ( X d ' X e )cos( o )}
I qo ( X q X d' ){( X e X q )sin( o ) Re cos( o )}] Including the effect of a modern solid state exciter, the
where, 1/ Ki Re ( X q X e )( X d X e ) (6)
2 ' machine model in terms of these constants is given
below:
A similar model has been developed for the type 1 of
K2 Ki [{Re . Eqao I qo {Re2 ( X e X q )2 }] (7)
IEEE by Anderson and Fouad [7]. The mathematical
models based on constants K1 to K6 are sufficiently
1/ K3 1 Ki ( X d X d' )( X q X e ) (8) accurate. The pertinent equations are given below:
K 4 Vb Ki ( X d X d' ){( X q X e )sin ( o ) K3 EFD KK
(9) Eq' = 3 4 ' (13)
Re cos( o )} 1 K3Tdo 1 K3Tdo
'

K5 ( KiVb / Vto )[ X d' Vqo{Re cos(o ) ( X q X e )sin ( o )} Te K1 K2 Eq' (14)

X qVdo{( X X e )cos(o ) Re sin (o )}] (10)
'
d
Vt K5 K6 Eq' (15)
K6 (Vqo / Vto )(1 Ki X ) (Vdo / Vto ) Ki X q Re
'
d (11)
131
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 10, October 2013)
j Tm Te -the linearised swing equation (16) Since the purpose of a PSS is to introduce a damping
torque component, a logical signal to use is the speed
deviation. Excitation control, governor control and PSS
V. EXTENSION T O MULTIMACHINE SYSTEMS
can be included in detailed modeling of the set.
The approach of one machine on infinite bus holds
good only for machines or group of machines the total VII. C ASE S TUDIES ON DYNAMIC STABILITY
generating capacity of which is insignificantly small with
Separate case studies have been made on dynamic
respect to the grid power. In such cases the changes in
stability of 210 MW turbo generator sets based on flux-
the operating conditions of the generator(s) does not
linkage and current state space models, for different
appreciably affect the grid system. However, if the
conditions of loading. The effect of exciter has been
capacity of the generator(s) is comparable to the grid
neglected in this study.The stability has been predicted
power then the approximation is no longer valid. Then all
from the eigenvalues of the state matrix. Characteristic
the synchronously coupled machines and the network
equation of the matrix has been found out by Fadeev-
must be modeled separately. To find out initial
Leverrier method and the eigenvalues have been obtained
conditions, it is essential to make a load flow study
by using Lin-Bairstows method [14][15]. MATLAB-
beforehand. Then these are to be recast to a common
tools have been used for check up [16].
reference frame. It is expedient to reduce the number of
Another case study has been made on the same
nodes and the number of machines for dynamic stability
machine, based on reduced order model, based on the
analysis to get a faster solution. This is made by the
constants of Concordia and De Mello. In this case the
techniques of network reduction and equivalencing. It
exciter has been modeled according to IEEE 1s (solid
has also been suggested by Anderson and Fouad to
state excitation). The system stability has been predicted
represent the machine(s) near to the location of
from the characteristic equation of the state matrix using
perturbation in details and the remote machines
Rouths array.
approximately, thereby making a further saving of
computer run-time. Inter-area oscillatory modes are A. Case-1 (under continuous maximum rating)
important in multi-machine systems- they have to be The first case-study has been made on a 210 MW
damped out using power system stabiliser [3][4][12]. turbo generator set connected to infinite bus through
the generator-transformer. The machine delivers its rated
VI. P OWER SYSTEM S TABILIZER FOR OPTIMIZING output at rated (lagging) power factor as is characteristic
P ERFORMANCE of peak hours of evening. Boundary conditions are given
Power system stabilizers are extensively used in at generator node. The flux-linkage state space model has
modern power system as effective means of enhancing been used to represent the machine. Stability has been
overall system stability [4][13]. It extends the system examined by applying Rouths criterion. The boundary
stability limit by modulating governor excitation so as to conditions are given at generator terminals. All
provide damping of the oscillations of synchronous parameters are in p.u. and the time constants in sec.
machine rotors relative to one another. It produces a
A1. Machine Parameters
component of electrical torque on the rotor in phase with
'
speed variation. The implementation details differ Ra = 0.001096 ; X a = 0.152 ; X d = 2.225 ; X d =
depending upon the input signal employed. However, for 0.305 X d" = 0.214 ; X q = 2.11 ; X q" = 0.311
any input signal the transfer function of the PSS must
compensate for the gain and phase characteristics of the Td' = 0.96 ; Tdo' = 7 ; Td" = 0.121 ; Tdo" = 0.1725 ;
excitation system, generator and the power system, which Tq" = 0.08844 ; Tqo" = 0.6 ; H =4.18 ; D = 1.0
collectively determine its transfer function.
Two of the commonly used types of PSS are the X e (referred to rated MVA of the generator ) = 0.1366 ;
type and T type. The structure in both the Re : neglected
type is same, irrespective of the input signal. It comprises
a pair of lead-lag network in cascade, which provides the A2. Operating Conditions
desired phase lead and a gain for compensation. The The generator power: P = 0.85p.u. at 0.85 lagging p.f.
basic function of a power system stabilizer is to add The generator terminal voltage: Vg 1 ;
damping to the generator rotor using auxiliary
stabilising signals. The stabilizer must produce a The infinite bus voltage: Vinf = 0.9307 7.003o
component of electrical torque in phase with the Power angle: 40.307o ;
deviation in rotor speed.

132
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 10, October 2013)
D-Q axes components of the voltage and current: TABLE-3
(Rouths tabulation)
Vd = -0.6469 ; Vq = 0.7626 ; I d 0.9516 ; 1 1.004536 0.00232515 0.000000846
s7
I q 0.3074 ; E E fd 2.8804 ; I fd 2.4067 2933

s6 0.108927200 0.07399189 0.000061547 0.000000000

Flux-linkages: 0000 33 8128
D-axis armature linkage= 1.3218 ; 0.325257300 0.001760118 0.000000838
s5
Field flux linkage= 1.9700 ; 0000 8317
Damper flux linkage= 1.5724
s4 0.07340244 0.000061266 0.000000000
Q-axis armature linkage= 1.1236 ; 41 8128
Q-axis damper flux linkage= 1.0426
s3 0.001488637 0.000000835
Infinite bus constant: K = 1.6121; 0000 2302
Its axis components: Kd = 1.1849 ; Kq 1.0930
s2 0.000020082 0.000000000
TABLE-2 4800 8128
(Elements of State-Matrix A(7,7)*1000) 0.000000774
s1
9825
-16.43 9.441 6.163 -878.4 246.3 -730.2 667.2
s0 0.000000000
8128
1.7833 -4.515 2.6013 0 0 0 0
The system is stable as all elements of the 1st. column
12.79 11.77 -25.49 0 0 0 0
are positive.
1138.4 -166.4 -263.9 1.1712 -3.708 762.04 823.34 B. Case-2 (Leading p.f. Condition)
The next case-study has been made on the same
0 0 0 59.08 -63.66 0 0
machine for slightly leading power factor conditions as is
characteristic of lean period of night. This time current
0.5988 -0.25 -0.396 - 0.749 0.4956 0 0
state space form has been used.
0 0 0 0 0 1000.0 0 B1. Operating Variables And Initial Conditions-
i. Rated power P =0.85 at 0.985 leading power factor.
Coefficients of the characteristic equation (Fadeev- Boundary conditions given at generator node
Leverrier method):
The generator terminal voltage: Vg = 1.05 0o;
s 7 1.1089s 6 1.0045s5 7.399e 2s 4 2.325e 3s3 The infinite bus voltage: Vinf = 1.0711 5.9638o ;
6.1547e 5s 2 8.463e 7s 8.1277e 10 0 Power angle: = 66.23o;
Stability analysis has been made in this case by Axis components of voltage and current:
applying Routh's criterion to the characteristic equation. Vd = -0.961 ; Vq = 0.4232
The Routh's tabulation is given below: I d = -0.6837 ; I q = 0.456 ;
Induced emf and field current:
E EFD 1.9453 ; I FD 1.6254
D-axis/Q-axis armature linkage = 0.7344/ 1.6667;
Infinite bus constant, K =1.8552 ;
Its d-axis/q-axis components, K d , K q : -1.7664/ 0.5672

133
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 10, October 2013)
TABLE-3 TABLE-4
(Elements of State-Matrix A(7,7)*1000) (Elements of State-Matrix A(7,7)*1000)

-19.65 1.089 8.055 -6408 -6408 -5062 1618 -19.65 1.089 8.055 -6408 -5585 -3802 3603

7.376 -4.256 14.072 2405.0 2096.1 1899.7 -607.2 7.376 -4.256 14.072 2405.0 2096.1 1426.9 -1352

11.688 3.015 -23.25 3811 3321.6 3010.4 -962.3 11.688 3.015 -23.25 3811.2 3321.6 2261.2 -2143

5276.1 4631.4 4631.4 -15.39 20.523 1279.4 3946.4 5276.1 4631.4 4631.4 -15.39 20.523 2831.4 2971.9

-4848 -4255 -4255 14.143 -23.55 -1176 -3626 -4848 -4255 -4255 14.143 -23.55 -2602 -2731

-3.622 -65.29 -65.29 -128.9 -92.45 -119.6 0 -2.72 -49.04 -49.04 -85.75 -30.69 -119.6 0

0 0 0 0 0 1000.0 0 0 0 0 0 0 1000.0 0

Coefficients of characteristic polynomial (found out by Coefficients of characteristic polynomial (by Fadeev-
Fadeev-Leverrier method) are given below: Leverrier method) are given below:

s 7 0.2057 s 6 + 1.371 s 5 + 0.1928 s 4 + 0.3656 s 3 + s 7 0.2057 s 6 + 1.346 s5 0.1903 s 4 0.3406s 3

2
1.598E-02 s + 1.776E-04 s -1.793E-07=0 1.336 E 02s 2 1.279E 04s 7.743E 07s 0

As all the coefficients of the characteristic polynomial The real roots by N-R method: s(1) =-2.949653e-02
are not positive, the system is unstable. This is confirmed No more real roots- 3 pairs of complex roots are being
by finding out the real root of the polynomial by N-R found out by Lin-Bairstow's method:
method: s = 9.302644E-04. It is a pole in the right half
of the s-plane showing instability of the system at the Complex pair(1): 5.1354e 03 j 7.216e 03
specified operating condition. frequency of oscillation= 0.3608 Hz.;
damping ratio= 0.6198
ii. Half rated power P =0.425 at 0.985 leading power
factor. Boundary conditions given at generator node Complex pair(2): 6.1265e 02 j 0.5755
The generator terminal voltage: Vg = 1.05 0o; frequency of oscillation= 28.78 Hz.;
damping ratio= 5.196e-02
The infinite bus voltage: Vinf = 1.0591 3.0116o ; Complex pair(3): 2.1709e 02 j 0.9992
Power angle: = 43.46o; frequency of oscillation= 49.95Hz.;
Axis components of voltage and current: dampingratio= 0.1466
Vd = -0.7223 ; Vq = 0.7621
As all the eigenvalues are either negative real or
I d = -0.227 ; I q = 0.3425 ; complex with negative real components, the system is
Induced emf and field current: stable.
E EFD 1.2678 ; I FD 1.0592 C. Case-3(Based on Reduced Order Model, Including
D-axis/Q-axis armature linkage = 1.321/ 1.2519; Exciter)
Infinite bus constant, K =1.8345 ; Now, the current state space model is being worked
Its D & Q-axes components- K d , K q :-1.3302 /1.2632 out in terms of Concordia and D'melloconstants.
Saturation and the effect of exciter have been included in
the study.

134
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 10, October 2013)
C1. Machine parameters: Table- VI
' (The Routh's tabulation)
X ds = 1.7 ; X qs = 1.64 ; X ds = 0.245 ; Z e = 0.02 + j 0.4;
other parameters are unchanged. s5 1.0 0.02146044 0.000003341687

Exciter constants (IEEE model 1s, with ref. to fig. 2):

s4 0.5099791 0.000158290 0.0000000095924
K a = 300 Ta = 314 Kf = 1 T f = 47.1

Operating variables and initial conditions: s3 0.02115005 0.000003322877

The generator power, P = 1.0 p.u. at a lagging power
factor of 0.85. s2 0.00007816764 0.0000000095924

Vt = 1.0 0o; Vb = 1.0106 3.1224o ; = 45.51o;

s1 0.0000007274391
Vd = -0.63062 ; Vq = 0.77609
I d = -1.1116 ; I q = 0.38527; Eqa = 2.5995 s0 0.0000000095924
Vb = 0.82838 -27.9 o

Changes of sign in the first column = 0. The system is

The Concordia &D'mello Constants for this condition stable for the given initial conditions.
of loading: The eigenvalues are being found out by MATLAB-
K1 = 1.0751 K 2 = 1.2578 K 3 = 0.3072 tools:
K 4 = 1.7124 K 5 = -0.0477 K 6 = 0.4971 s1 = -0.4645 ; s2 = -0.0415 ; s3 = -0.0032 ; s4 , s5
Table-V = -0.0004 j 0.0126
(State matrix for the reduced order model)
All the real eigenvalues are negative and the complex
-0.46502 0 -0.24462 0 -0.14286
eigenvalues have negative real part. So the system is
stable under given initial conditions.
-0.00048 -0.00026 -0.00041 0 0
VIII. CONCLUSION
0 1 0 0 0 Dynamic stability is a kind of steady state stability. It
applies to operating conditions of a machine or a group
0 0 0 -0.04152 -0.00007
of machines, subjected to small perturbations. Both
oscillatory and exponential modes are created under
0 0 0 -0.00342 -0.00318
small impact. Their nature depends on the machine
The corresponding B-matrix:
parameters and the initial conditions. If there be a
growing oscillatory or exponential response, then the
0 0.00038 0 0 0 system is unstable. This phenomenon occurs if the
eigenvalues of the state matrix are either positive real or
The characteristic equation is given below: they are complex with positive real parts i.e. situated in
the right half of s-plane. The state matrix is found out
s 5 + 0.51 s 4 + 0.02146 s 3 + 1.583E-04 s 2 + 3.342 s + either in terms of current states or flux-linkage states by
9.59E-09 = 0 applying linearizing techniques around the operating
point. The model can be elaborated to include the effects
Stability analysis has been made by using Routh's
of saturation, exciter and governor control and PSS.
criterion on the characteristic equation.

135
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 10, October 2013)
Sometimes reduced order models are also used to [2] Klein M., Rogers G.J., Moorty S. and Kundur P., Analytical
investigation of factors influencing power system stabilizer
reduce the dimension of the state matrix and run-time of
performance, paper 92WM016-6EC, presented at the IEE PES
the computer. The idealization of one machine on infinite Winter meeting, New York, January, 1992.
bus is made if the rating of the machine is small [3] Klein M., Rogers G.J., and Kundur P., A fundamental study of
compared to the total capacity of the grid. Otherwise interarea oscillations, IEEE Trans, Vol. PWRS-6, No. 3, pp-914-
multimachine dynamic stability analysis has to be made. 921, August, 1991.
If the addition of fast excitation controller gives rise to [4] Kundur P., Lee D.C., Zein EI-Din H.M., Power system
poor or negative damping, then a variable structure stabilisers for thermal units: analytical techniques and on-site
validation, IEEE Trans., Vol. PAS-100, pp. 81-96, January,
Power System Stabilizer (PSS) should be added. 1981.
Dynamic stability analysis has been made in this paper [5] Kundur P., and Dandeno P.L., Practical application of eigenvalue
for 210 MW turbo generator set using both flux-linkage techniques in the analysis of power system dynamic stability
and current state space models for different conditions of problems, Proc. of 5th. powers system computation conference,
loading. While the machine delivers rated power at rated Cambridge, England, September, 1975.
p.f., it is found to be stable. But when the power factor is [6] Radhakrishna C., Stability studies of AC/DC power systems,
Ph. D. Thesis, Indian Institute of Technology, Kanpur, India,
slightly leading (0.985), the generator becomes 1980.
dynamically unstable. But under light load condition (0.5 [7] Anderson P.M. and Fouad A.A., Power System Control and
p.u. loading) the generator becomes dynamically stable. Stability- IOWA State university, Galgotia Publications.
In each case, the machine has been idealized as one [8] Singh L.P., Advanced power system analysis and dynamics,
machine on infinite bus and the effect of exciter has been New Age Int. 3rd. Ed., 1996
neglected. [9] Kundur P., Power system stability and control, The EPRI Power
The other case study has been made on the same System Engineering Series, McGraw-Hill, Inc., 1994.
machine. This time the effect of saturation and the solid [10] Farmer R.G., Power system dynamics and stability, Book,
Arizona state university.
state exciter have been included. A reduced order model
[11] Adkins B. and Harley R.G., Generalized theory of a.c.
based on the constants K1 K6 of Concordia and machines, Book, Chapman & Hall
DMello has been made. The exciter has been [12] Venikov V., Transient processes in electric power system,
represented according to IEEE model 1s. Dynamic Book, MIR Publishers.
stability study has been made for operation at continuous [13] Kundur P., Power system stability and control, The EPRI power
maximum rating (CMR), using Rouths array on the system engineering series, McGraw Hill.
characteristic equation of the state matrix. It is found that [14] Kreyszig E., Advanced engineering mathematics, Wiley, New
York, 2001.
the machine at this operating condition is dynamically
[15] Rajaraman V., Computer-oriented numerical methods, PHI,
stable. 1995
REFERENCES [16] Pratap R., Getting strated with MATLAB7, Oxford, Indian Ed.
[1] Watanabe M., Mitani Y., Iki H., Uriu Y., and Urano Y.,Dynamic
stability assessment of customers power systems with detailed
models implementation, 17th Power Systems Computation
Conference Stockholm Sweden - August 22-26, 2011

136