ISSN (Print) : 0974-6846

Indian Journal of Science and Technology, Vol 7(S6), 4047, October 2014 ISSN (Online) : 0974-5645

Multi-Machine Small Signal Stability Analysis For

Large Scale Power System
C. Chennakesavan* and P. Nalandha
Department of Electrical and Electronics Engineering, AMET University, Chennai;
cckesavan@gmail.com; nalanpsel@gmail.com

Abstract
The main objective of this paper is a fundamental study of the nature of inter-area oscillations in power systems. Small-signal
stability refers to the ability of a power system to maintain synchronism under small disturbances Instability may arise
in two forms increase of rotor angle due to lack of sufficient synchronization torque and rotor oscillations of increasing
amplitude due to lack of sufficient damping torque. To study small signal stability analysis synchronous machine model,
transmission line model and two area system model and the dynamic state matrix eigenvalues and eigen vectors are con-
structed and the small signal stability analysis done with the developed algorithm. The 11 bus systems are considered here
for study of oscillations. The effects of the system structure, generator modelling, excitation type, and system loads are dis-
cussed in detail. In the study, only small signal stability analyses are used to determine the characteristics of the system. In
the power system leads to the development of many oscillations at low frequency in the power system. This paper presents
the effect of the load model it was easier to identify unstable modes of oscillation.

Keywords: Power System Modeling, Eigen Value Analysis, Participation Factors, Small Signal Stability, TYPE-2B MODEL

1. Introduction Small signal stability is the ability of the system to

With the increasing electric power demand, power
systems can reach stressed conditions, resulting in unde-
sirable voltage and frequency conditions1. The eigen value
analysis was done for small signal stability of 11 bus 2 area
system. The drawback of eigenvalue analysis is that the
higher order terms neglected from the liniearised set of
equation, Here particularly in a system is described by
strongly non-linear equations.
Modern electrical power systems have grown to a
large complexity due to increasing interconnections,
installation of large generating units and extra-high volt-
age tie-lines etc.
The system response to such disturbances involves
large excursions of generator rotor angles, power flows,
bus voltages, and other system variables2. It is important
that, while steady-state stability is a function only of
operating conditions.
maintain synchronism under small disturbances which
occur continually on the system due to small variations
in loads and generation or other small disturbances on
the system4. A disturbance is considered to be small if
the equations that represent the dynamic performance
of the system can be linearized for the purpose of anal-
ysis function of both the operating conditions and the
disturbance.
In this paper a new method for calculation of the two
area four generator system is investigated. Stability stud-
ies for power system planning, operation and control
rely immensely on computer based power system simu-
lation tools. Simulation tools use mathematical models
that predict the dynamic performance of the system7.
It is crucial that these power system models be mod-
eled accurately to predict the actual performance of the
system.

*Author for correspondence


2.System Description and
Modelling
A standard IEEE 11 bus system1 as shown in Figure 1.
2.1 Synchronous Machine Model
The synchronous generators are modelled as classical
machines with and as state variables. The generator
is represented by the Norton equivalent for network solu-
tion. The admittance of the generator is included in the
main diagonals of Y matrix.
dw
J = Tm - Te  (1)
dt
dd
= w - w0  (2)
dt
2.2 Transmission Line Model (Short Line)
A transmission line is modeled as a short line. It consists
of a series impedance comprised of a resistance R and
reactance X between the terminal nodes.
2.3 Load Model
The loads are modeled as a constant admittance in
transient stability studies.
2.4 Introduction of Two-Area System
The machines are considered to be TYPE -2B Model. The
system contains eleven buses and two areas, connected
by a weak tie between bus 7 and 9. Totally two loads are
applied to the system at bus 7 and 9. Two shunt capacitors
Figure 1. IEEE two area 11 Bus System
Vol 7 (S6) | October 2014 | www.indjst.org
C. Chennakesavan and P. Nalandha
are also connected to bus 7 and 9 as shown in the figure 1
below. The system has the fundamental frequency 60 Hz.
The system comprises two similar areas connected
by a weak tie3. Each area consists of two generators, each
having a rating of 900 MVA and 20 kV. We will perform:
Power-flow calculation
Linear Analysis and Modal Analysis
Time-domain simulation.
The left half of the system is identified as area 1 and the
right half is identified as area 2. The saturation of the syn-
chronous machines are not identical.
This chapter describes the methodology for incor-
poration of the model for small-signal analysis of single
machine-infinite bus and multi-machine configurations.
This is followed by relevant results obtained for these
configurations.
3. Small Signal Stability
Small Signal Stability is the ability of the power system to
maintain Synchronism under small perturbations. Such
perturbations occur continuously on the system because
of small variations in load and generation where the sys-
tem continuously adjusts itself to changing conditions.
Restoring forces acting on the machine help them
to maintain stable conditions. The system must be able
to operate satisfactory under these conditions and suc-
cessfully supply the maximum load. The Small Signal
disturbances are considered suffiently small for lineariza-
tion of system equations to be permissible for the purpose
of analysis.
The most common form of synchronism, monotoni-
cally, in the first few seconds following the fault due to lack
of synchronizing torque and damping torque. The stabil-
ity of the following types of oscillations is of concern.
Local modes are associated with the oscillations of
generating units at a particular station with respect to the
rest of system. These oscillations are localized in a small
part of the power system.
Inter area modes are associated with the oscillations
of many machines in one part of the system against
machines in the other parts7.
3.1 State Equation
For a classical model of the synchronous machine the
states are incremental changes in rotor speed and angle.
Indian Journal of Science and Technology 41
 Multi-Machine Small Signal Stability Analysis For Large Scale Power System

In our analysis the super bar notation on the per unit The state variables for the machine are
quantities are dropped out. The state equations for the Dw Dd DEq DEd DEd DEd. The right hand side of the

SMIB system shown in the Figure 2 are given by differential equations for the machine and the controllers
will contain algebraic variables such as current, voltage or
D.w - K D -K S 1 power and incremental changes in them should be elimi-
= 2H Dw
2H + 2H DTm  (3) nated to obtain the equations in state variables canonical
. Dd
Dd w 0 0 0 form. Once the equations in the state variables canonical
form are obtained, we can apply eigen value technique to
This is the form of X = Ax + Bu. The elements of the
access stability.
state matrix A are seen to be dependent on the system
parameters KD, H, XT and the initial operating condition K D T 0 0 0 0
T m
represented by the values of E1 and 0. .w 2H
e

Where, ws 0 0 0 0 0 0
is in per unit . w 1
d 0 1 1 0 0 0 d EFD
is in radians . Td0 Td0 Td0
The synchronizing torque coefficient is given by Eq 1
1 1 Eq
= 0 0 0 0 0 + E (5)
E Tq0
. Ed Tq0
FD

E V Tq0
KS = cos d  (4) d E 1
( Xd + Xtr + Xline ) . 0 1 0 0 1 0 d EFD
Eq Td0 Tq0 Ed Td0
. 1
3.2 Multimachine Model Ed 0 1 0 0 0 1 EFD
Td0 Tq0 Tq0
The base system is symmetric; it consists of two identical
areas connected through a relatively weak tie. Each area Assumptions:
including two generating units with equal outputs. The
full symmetry of the base system clarifies the effect that The machines are considered to be classical (no

various factors have on the inter-area mode. controllers)
Dynamic data for the generator and excitation systems Damping ignored

used in the study are given in Appendix-A. In setting up Loads are assumed as constant a dmittances

the various power flows used in the studies, capacitors
were added as necessary to ensure that the systems volt- Preparation:
age profile was satisfactory. The initial conditions for , and voltages are obtained
The electro-mechanical modes of oscillation are pres- from load flow and the past history terms for and are
ent in this system; two inter-area modes, one in each area, obtained from the initial conditions.
and one inter area low frequency mode, in which the gen-
erating units in one area oscillate against those in other
4.Multi-Machine Small Signal
area.
Our experience with large inter connected systems Stability Algorithm for
confirms some of the results of our studies using the Type-2b Model
small system, and we are confident that the general
1. Taken IEEE 11 bus 2 area 4 machine system
conclusions drawn from our work, will apply to large
2. Formulate the admittance matrix
systems. The following equations pertain to the four-
machine system. YV = I (6)
3. After that reduced the network equations
3.3 State Space Model red -1
YGG = YGG - YG , NG * YNG , NG * YNG , G  (7)
Our aim in this section is to derive the equations for
multi machine system that represent the dynamics of 4. Initialize the power, current, voltage in transient
the machine and the controller in the state space form. condition & sub-transient condition etc.

42 Vol 7 (S6) | October 2014 | www.indjst.org Indian Journal of Science and Technology
 C. Chennakesavan and P. Nalandha

5.Find out the transformation matrix & inverse 12. Finally our state variables are
transformation matrix Eq Ed Ed Ed  (21)

e jq1 0 0 0 e - jq1 0 0 0
13. Find out the eigen values from above state variables
jq 2 - jq 2
0 e 0 0 0 e 0 0
T= jq 3
inv T = - jq 3

0

0 e 0
jq 4

0

0 e 0
- jq 4
5. Eigen Analysis
0 0 0 e 0 0 0 e
With the linearized power system model in matrix form
 (8) shown in equation the system can now be analyzed through
6.Then the reduced equations into individual machine the use of eigenvalues and eigenvectors. To examine the
rotor coordinates. free response of the system the inputs are put to zero.

M = inv(T) YGGred T (9) x = [A] x (22)

7.After applying the linearization of network equations where x is a state vector and A is the state matrix of size
in individual machine rotor coordinates. n x n;
The state equation given by is further analyzed by tak-
I = I0 + DI  (10) ing the Laplace transform. The new equation derived in
E = E0 + DE  (11) the s domain is given by

M = M0 + DM  (12) det (s[I] [A]) = 0 (23)

The values of s that satisfy (poles of system) are known
T = T0 + DT  (13)
as eigenvalues of the matrix A. The eigenvalues may be
8. Small change in current is real or complex. For an n x n matrix they are n eigenval-
ues. If A is real eigenvalues occur in complex conjugates
DI = [M0 DE - j(DdM0 - M0 Dd)E 0 ]  (14)
of the form [6]
9. Linearization of differential equations: = j. (24)
The mechanical side differential equations are given by
dDw i DTmi - DTei - K D Dw i 5.1 Eigenvalues and Stability
=  (15)
dt 2H i The power system is stable if all of the eigenvalues are on
the left-hand side of the imaginary axis of the complex
10. Linearization of swing equation is
plane; otherwise it is unstable. If any of the eigenvalues
dDd i comes over into the right hand side of the imaginary axis,
= w s Dw i  (16)
dt the corresponding modes are said to be unstable, and so
is the system.
11.We illustrate how to include the damper windings in
The desired aim then is to have all eigenvalues in the
the formulation of state variables2.
left hand plane. This stability is confirmed by looking at
. 1 the time dependent characteristic of the oscillatory modes
DE q = - [DEq + (X d - X d )DId - DEFD ]  (17)
Tq0 corresponding to each eigenvalue given by e t.
A real eigenvalue corresponds to a non oscillatory
. 1 mode. A negative real eigenvalue indicates a mode that
DE d = - [DEd + (X q - X q )DIq ] (18)
Tq0 decays in time (the larger the magnitude of the eigen-
value the quicker the decay)9. A positive real eigen value
. 1 indicates a mode that grows with time and is system will
DE q = - [DEq + (X q - X q)DId - DEq ]  (19)
Tq0 experience aperiodic instability.
A conjugate pair complex eigenvalues indicate oscilla-
. 1 tory modes of response.
DE d = - [DEd + (X d - X d)DId - DEd ]  (20)
Td0 = j (25)

Vol 7 (S6) | October 2014 | www.indjst.org Indian Journal of Science and Technology 43
 Multi-Machine Small Signal Stability Analysis For Large Scale Power System

1. If a conjugate of pair of complex eigenvalues has nega- i [A] = i

tive real parts this corresponds to an oscillatory
mode that decays with time and the system is said to Y i1
be globally stable. Y
i2
2. If a pair has positive real parts, the corresponding Yi =  (29)
oscillatory mode grows exponentially with time and
eventually dominates the system behavior. Such a sys-
Y in
tem is said to be unstable.
3. If any one of the eigenvalues has a real part, the system The left eigenvectors, together with the initial con-
will have an undamped oscillatory response. ditions of the system state vector x, determine the
magnitudes of the modes.
Eigenvalues associated with an unstable or poorly damped
oscillatory mode are called dominant modes since 5.3 Mode Shape
their contribution dominates the time response of the
The right eigenvector gives the mode shape.
system.
The mode shape is the relative activity of the state vari-
The real component of an eigen value gives the
ables when a particular mode is excited. Thus the degrees
damping, and the imaginary part gives the frequency of
of activity of the Kth state variable xk in the Ith mode is
oscillation.
given by the element ki if the right eigenvector i.
Frequency of oscillation (Hz): The magnitude of the elements if i gives the extent
of the activities of the n state variables in the Ith mode,
w
f =  (26) and the angles of the elements give phase displacement of
2p
the state variables with regard to the mode. Thus we can
The damping ratio: use the mode shape to analyze the magnitude and phase
displacement of the speed and rotor angle state variables
-s in an oscillatory mode.
z=  (27)
s 2 +w2
6. Participation Factors
5.2 Eigen Vectors The participation factor P is useful in identifying those
Eigen analysis also includes the computation of eigen- states which have the most influence on any mode. The
vectors. For any eigen value i, the column vector i that participation factor is non-dimensional.
satisfies is called the right eigenvector for i P1i j1iY i1
P j Y
2i i 2
Pi =
2i
[A]i = i i = 1, 2, n  (30)


j1i Pni jniY in
j
2i The element pki = ki ik is called a participation
ji =  (28)
f actor. It is a measure of the relative participation of the

kth state variable in the ith mode, and vice versa5. In effect
j 2i participation factors are useful in identifying those state
variables which have the most influence on any mode.
The right eigenvector show the distribution of the The higher the value of participation factor of a state
modes of response (eigenvalues) through the power sys- for a corresponding mode, the more active that state is in
tem state variables. Correspondingly, there is a row vector that mode when compared with the other states.
i that satisfies (24) and is called the left eigenvector of Thus the values of participation factors can reveal
[A] associated with the eigen value i. which generators are involved in a particular mode. It can

44 Vol 7 (S6) | October 2014 | www.indjst.org Indian Journal of Science and Technology
 C. Chennakesavan and P. Nalandha

reveal which machine or machines could go out of step Eigen-analysis of the system can then be performed by
for any known mode or modes that might cause problem obtaining the eigenvalues of the state matrix. The eigen-
in the power system in the advent of load variation10. values are then check for stability.
The participation factors are used to identify the areas
in the power system where any mode or (oscillation) has 6.3 Load Model Procedure
most of its effect.
1. The type of load at various buses was selected i.e.
changing parameters np and nq
6.1 Eigen Value Sensitivity to Load Changes 2. The system matrix A sys was computed
The sensitivity of an eigen value i to an element kj of the 3. The eigenvalues are computed and check for stability
state matrix is equal to the product of the left eigenvector analysis.
element ki and right eigenvector element ji.
7. Simulation Results
li
= Y i k ji j  (31)
a k j A standard 11-bus 4 machine 2 area as shown in system.
The initial relative rotor angle obtained from the load flow
In this analysis the load power at a bus in the power solution or initial condition.
system was varied and the system state matrix computed.
Table 1.
The elements kjC of the state matrix that showed the
EIGEN VALUES Damping
greatest changes due to the load change were identified. Frequency (HZ)
Thus the sensitivity of the eigenvalues to these elements NO Real Imaginary Ratio
of the state matrix A sys can be computed using equation. 1 0.76e 3 0.22e 2
0.0003 0.331
In computing the eigen value sensitivity the elements 2 0.22e 2
that show a large variation with a change in load give 3 0.096
an indication of the sensitivity of the eigen value to load 4 0.111 3.43
changes. 0.545 0.032
5 0.111 3.43
6 0.117
6.2 Simulation 7 0.265
The test system used in this investigation is the IEEE 8 0.276
11 bus test system. Matlab 7.5 was used to analyze this 9 0.492 6.82
system. The system is a consists of 4 generators. Software 1.087 0.072
10 0.492 6.82
was written to perform the small signal analysis on the
11 0.506 7.02
system. This program follows three main algorithms: 1.117 0.072
12 0.506 7.02
Loads system data 13 3.428
Performs load flow of system, calculates initial 14 4.139
condition, linearize system equation and construct the 15 5.287
matrices and finally the system state matrix A sys.
16 5.303
Calculate systems eigenvalues a check for unstable
17 31.03
mode and gets the participation and mode shape for
critical Eigenvalues. 18 32.45
19 34.07
The effect of loading was analyzed by increasing the load 20 35.53
at a particular bus or buses. At each change in loading 21 37.89 0.142
the initial conditions of the state variables were calculated 0.023 1.0
22 37.89 0.142
after running a load flow [9]. Next, linearization of the
23 38.01 0.038
system equation was done and the state matrix A sys is 0.006 1.0
24 38.01 0.038
formed.

Vol 7 (S6) | October 2014 | www.indjst.org Indian Journal of Science and Technology 45
 Multi-Machine Small Signal Stability Analysis For Large Scale Power System

Figure 2. (a,b,c) Mode shapes of rotor angle

From this table 1 all eigen values are coming negative. Bus 7 PL = 967MW QL = 100 MWAR QC = 200 MWAR
So we see that system is Stable. There are three rotor angle Bus 9 PL = 1,767MW QL = 100 MWAR QC = 350 MWAR
modes of oscillation. Their mode shapes are shown in
Figure 2(a, b, c)
9. Appendix
8. Conclusion The system consists of two similar areas connected by a
weak tie. Each area consists of two coupled units, each
In this investigation a dynamic model for the power sys- having a 900MVA & 20 KV. The generator parameters in
tem was developed to investigate the development of low per unit on the rated MVA and KV base are as follows.
frequency oscillations in a power system. The test system
was used the IEEE 11 bus test system. A power system Xd = 1.8 Xq = 1.7 Xl = 0.2 Xd = 0.3
model was enveloped in Matlab and testing was done to Xd = 0.25 Xq = 0.25 = 8.0 S
Td0 = 0.4 SS
Tq0
determine the power system stability. = 0.03 S
Td0 Asat = 0.015 Bsat = 9.6
= 0.05 S
Tq0
Software developed performed the following function:
Xq = 0.55 Ra = 0.0025 jT1 = 0.9 kD = 0
load flow of power system, calculation of the initial values
of the power system model and the construction of the H = 6.5 (for G! & G2) H = 6.175 (for G3 & G4)
system state matrix. The software then applied the linear
Each Step-up transformer has an impedance of
analysis tools to find Eigenvalues of A sys, Participation
0 + j0.15 per unit on 900 MVA and 20/230 KV base, and
factors, and mode shape.
has an off-nominal ratio of 1.0
The Eigenvalues of A sys, Participation factors, and
The parameters lines in per unit on 900 MVA &
Mode shape can be used to determine if system could
230KV
develop oscillations and where they occur in the power
system if excited. R = 0.0001 pu/km XL = 0.001 pu/km Bc = 0.00175 pu/km
The purpose of this research was to investigate the
Contribution of load to low frequency oscillations in Then the generating units loaded as given below.
Power System8. As a result the system was analysis by first
G1 P = 700 MW Q=185MW Et = 1.03 20.2 G1
examining the effect of loading on the development of
G2 P = 700 MW Q = 235MW Et = 1.01 10.5 G2
system oscillations. Secondly the effect of load model on

low frequency oscillation was determined. G3 P = 719 MW Q = 176MW Et = 1.03 - 6.8 G3
In the power system leads to the development of G4 P = 700 MW Q = 202MW Et = 1.01 - 17.0 G4
many oscillations at low frequency in the power system.
Finally the effect of the load model was that it was easier The load and reactive power (Qc) by the shunt capacitors
to identify unstable modes of oscillation. at b uses 7 and 9 as follows

46 Vol 7 (S6) | October 2014 | www.indjst.org Indian Journal of Science and Technology

10. References
1. Kundur P. Klein M. Analytical Investigation of Factors
Influencing power System Stabilizers Performance, IEEE
Transaction on Energy, 1992.
2. Ramanujam R. Power System Dynamics, PHI Learning
Private Limited Conversion, 2009; 7(3).
3. Padiyar KR. Analysis of small signal voltage stability
in multimachine Systems using Detailed and Reduced
Formulations, IEEE Transaction on Electrical Power and
Energy Systems, 2003; 11(4).
4. Anderson PM. Fouad A. Power System Control and
Stability, The Iowa State University Press, Ames, Iowa. 1977.
5. Franken B and Anderson. Analysis of Converters
Connected to weak AC Systems, IEEE Transaction on
Power systems, 1990; 5(1).
Vol 7 (S6) | October 2014 | www.indjst.org
C. Chennakesavan and P. Nalandha
6. Kundur P. Power System Stability And Control, McGraw-
Hill, Inc. 1993.
7. M Ghandhari, G Andersson. Damping of inter-area and
local modes by the use of controllable components, IEEE
Transaction on Power Delivery, 2007; 10(4).
8. MA Pai. Energy, Function Analysis for Power System
Stability, Kluwer Academic Publishers, 1989.
9. Maurice H. Kent, Waynerfrancis a. Dynamic Modeling
of Loads in Stability Studies ieee Transactions on Power
Apparatus and Systems, 1996; (5) p. 88.
10. Elices A, L Rouco, H Bourles. Physical interpretation of
state feedback controllers to damp power system oscillations,
IEEE Transactions on Power System, 2004;19(1).
Indian Journal of Science and Technology 47