Analysis and Damping Control of Power System Low-Frequency Oscillations

Power Electronics and Power Systems
HaifengWang
WenjuanDu
Analysis and
Damping Control
of Power System
Low-frequency
Oscillations

Series editors
Joe H. Chow, Rensselaer Polytechnic Institute, Troy, New York, USA
Alex M. Stankovic, Tufts University, Medford, Massachusetts, USA
David Hill, The University of Hong Kong, Sydney, New South Wales, Australia
More information about this series at http://www.springer.com/series/6403
Haifeng Wang Wenjuan Du
Analysis and Damping

Control of Power System
Low-frequency Oscillations
123
Haifeng Wang
Beijing
China
Wenjuan Du
Beijing
China
ISSN 2196-3185
ISSN 2196-3193 (electronic)
ISBN 978-1-4899-7694-9
ISBN 978-1-4899-7696-3 (eBook)
DOI 10.1007/978-1-4899-7696-3
Library of Congress Control Number: 2016931424
Springer Science+Business Media New York 2016
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microlms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specic statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made.
Printed on acid-free paper
This Springer imprint is published by SpringerNature
The registered company is Springer Science+Business Media LLC New York
Preface
Power system electromechanical low-frequency oscillations threaten the stable

operation of power systems. A great effort has been spent by many researchers for
over half-century in order to understand and solve this complicated engineering
problem. There are two main methods for the analysis and damping control of
power system low-frequency oscillations. They are the damping torque analysis and
modal analysis based on the linearized models of power systems for the study of
small-signal angular stability. In this book, they are named as linearized methods or
small-signal methods. The damping torque analysis is simple and of clear physical
meaning. The phase compensation method developed on the basis of damping
torque analysis for the design of power system stabilizers (PSSs) has been widely
used in practice. The modal analysis is based on the modal control theory and has
been the mostly used method in large-scale multi-machine power systems. This
book is devoted to the introduction of those two linearized methods and their
applications in the design of conventional PSSs and more recently appeared FACTS
stabilizers. It is written as a textbook for postgraduate research students and a
reference book for power system researchers. In addition to the introduction of
modal analysis, great attention has been paid in the book to elaborate the principle
and applications of the damping torque analysis, which we have applied in our
research on the subject in recent 20 years.
We have used the main materials in this book in teaching and tutoring our
research students for years. Some examples in the book are the results of their
exercise work before starting their research projects. We would like to acknowledge
their contributions to the preparation of examples in the book. They are Yifu Lin,
Xiangfeng Wang, Yan Zhang, Shen Yang, Qianhui Wan, Ximin Li (Sect. 2.4), Chen
Lv (Sect. 3.3), Yi Ge (Sect. 4.4), Tianyu Su (Sect. 5.3), Zhijin Zhen (Sect. 6.3), and
Xiaobo Hu (Sect. 7.4.2).
Beijing
October 2015
Haifeng Wang
Wenjuan Du
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1 Power System Low-frequency Oscillations . . . . . .
1.2 Linearized Methods for the Analysis and Damping
Control of Power System Oscillations . . . . . . . . .
1.3 FACTS and Grid-Connected ESS . . . . . . . . . . . .
1.4 Controllers to Damp Power System Oscillations . .
1.5 Design of Damping Controllers to Suppress Power
System Oscillations . . . . . . . . . . . . . . . . . . . . . .
1.6 Organization of the Book . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
...........
...........
1
1
...........
...........
...........
3
5
7
...........
...........
...........
10
12
13
2 A Single-Machine Innite-Bus Power System Installed

with a Power System Stabilizer. . . . . . . . . . . . . . . . . . . .
2.1 Linearized Model of a Single-Machine Innite-Bus
Power System Installed with a Power System Stabilizer
2.1.1 General Linearized Mathematical Model. . . . . .
2.1.2 HeffronPhillips Model . . . . . . . . . . . . . . . . .
2.2 Modal Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.1 Basis of Modal Analysis Theory . . . . . . . . . . .
2.2.2 Applications of Modal Analysis . . . . . . . . . . .
2.3 Damping Torque Analysis . . . . . . . . . . . . . . . . . . . .
2.3.1 Damping Torque and Synchronizing Torque . . .
2.3.2 Damping Torque Analysis and Design
of PSS by Phase Compensation. . . . . . . . . . . .
2.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.1 Linearized Mathematical Models of an Example
System. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.2 Modal Analysis of Example Power System. . . .
2.4.3 Damping Torque Analysis of Example
Power System . . . . . . . . . . . . . . . . . . . . . . . .
........
17
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
17
17
26
32
32
37
42
42
...
...
47
53
...
...
53
59
........
66
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.....
.....
Power
.....
.....
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
vii
viii
Contents
2.4.4 Equivalence Between the Damping Torque

and Modal Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3 Damping Torque Analysis of Thyristor-Based FACTS
Stabilizers Installed in Single-Machine Innite-Bus
Power Systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1 A Single-Machine Innite-Bus Power System Installed
with an SVC Stabilizer . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.1 Extended HeffronPhillips Model
of a Single-Machine Innite-Bus Power System
Installed with an SVC Stabilizer . . . . . . . . . . . . . . . . .
3.1.2 Damping Torque Analysis of SVC
Stabilizer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
with a TCSC or TCPS Stabilizer . . . . . . . . . . . . . . . . . . . . . .
3.2.1 Extended HeffronPhillips Model
of a Single-Machine Innite-Bus Power System Installed
with a TCSC or TCPS Stabilizer . . . . . . . . . . . . . . . . .
3.2.2 Damping Torque Analysis of TCSC
and TCPS Stabilizers . . . . . . . . . . . . . . . . . . . . . . . . .
3.3 An Example Power System Installed with an SVC
Stabilizer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.1 Linearized Model of Example Power
System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.2 Design of SVC-Based Stabilizer . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4 Single-Machine Innite-Bus Power Systems Installed
with VSC-Based Stabilizers . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1 Damping Torque Analysis of a Shunt VSC-Based
Stabilizer Installed in a Single-Machine Innite-Bus
Power System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.1 Extended HeffronPhillips Model of a Single-Machine
Innite-Bus Power System Installed with a Shunt
VSC-Based Stabilizer. . . . . . . . . . . . . . . . . . . . . . . . .
4.1.2 Damping Torque Analysis of Shunt VSC-Based Stabilizer
Installed in Single-Machine Innite-Bus
Power System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2 Damping Function of a Stabilizer Added on a Static Synchronous
Series Compensator (SSSC) Installed in a Single-Machine
Innite-Bus Power System . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.1 Damping Torque Analysis of a SSSC Stabilizer Installed
in a Single-Machine Innite-Bus Power System . . . . . .
4.2.2 Design of a SSSC Stabilizer . . . . . . . . . . . . . . . . . . . .
73
79
..
81
..
81
..
81
..
90
..
97
..
97
. . 103
. . 107
. . 108
. . 114
. . 120
. . 121
. . 121
. . 121
. . 130
. . 133
. . 133
. . 139
Contents
4.3 Damping Function of a Unied Power Flow Controller (UPFC)

Installed in a Single-Machine Innite-Bus Power System . . . . .
4.3.1 Mathematical Model of a Single-Machine
Innite-Bus Power System Installed
with a UPFC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.2 Design of a UPFC Stabilizer Installed
in a Single-Machine Innite-Bus Power System . . . . . .
4.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.4.1 An Example Single-Machine Innite-Bus Power System
Installed with a BESS Stabilizer . . . . . . . . . . . . . . . . .
4.4.2 An Example Single-Machine Innite-Bus Power System
Installed with a UPFC Stabilizer . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5 A Multi-machine Power System Installed with Power System
Stabilizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1 Mathematical Model of a Multi-machine Power System Installed
with Power System Stabilizers. . . . . . . . . . . . . . . . . . . . . . . .
5.1.1 A Two-Machine Power System Installed
with Power System Stabilizers . . . . . . . . . . . . . . . . . .
5.1.2 A Multi-machine Power System Installed
with Power System Stabilizers . . . . . . . . . . . . . . . . . .
5.2 Modal Analysis and Control of Power System Oscillations
in a Multi-machine Power System Installed with Power System
Stabilizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.1 Eigensolution for the Analysis of Power System
Oscillations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.2 Design of Power System Stabilizers
in a Multi-machine Power System . . . . . . . . . . . . . . . .
5.2.3 Fixed Modes Associated with PSS Control . . . . . . . . . .
5.3 An Example Three-Machine Power System. . . . . . . . . . . . . . .
5.3.1 Example Power System and Its Linearized
HeffronPhillips Model . . . . . . . . . . . . . . . . . . . . . . .
5.3.2 Modal Analysis of Example Power System. . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6 Multi-machine Power System Installed with Thyristor-Based
FACTS Stabilizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
with Thyristor-Based FACTS Stabilizers. . . . . . . . . . . . . . . . .
6.1.1 HeffronPhillips Model . . . . . . . . . . . . . . . . . . . . . . .
6.1.2 General Linearized Model of an N-Machine
Power System Installed with Multiple
Thyristor-Based FACTS Stabilizers . . . . . . . . . . . . . . .
ix
. . 143
. . 143
. . 155
. . 163
. . 163
. . 172
. . 182
. . 183
. . 183
. . 183
. . 192
. . 199
. . 199
. . 204
. . 209
. . 214
. . 214
. . 221
. . 234
. . 235
. . 235
. . 235
. . 242
Contents
6.2 Analysis and Damping Control of Thyristor-Based FACTS

Stabilizers Installed in a Multi-machine Power System . . . . . . .
6.2.1 Damping Torque Analysis in a Multi-machine Power
System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.2 Selection of Installing Location and Feedback Signal of a
Stabilizer in a Multi-machine Power System . . . . . . . . .
6.2.3 Selection of Robust Installing Locations
and Feedback Signals of a Stabilizer by an
Eigensolution-Free Method . . . . . . . . . . . . . . . . . . . . .
6.2.4 Stabilizer Design in a Multi-machine Power System
Considering Robustness and Interaction of Stabilizers . .
6.3 An Example Two-Area Four-Machine Power System . . . . . . . .
6.3.1 Linearized Model . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.2 Selection of Installing Locations of Stabilizers . . . . . . .
6.4 Example Three-Machine Power System . . . . . . . . . . . . . . . . .
6.4.1 Dynamic Interactions Among PSSs Installed
in Example Three-Machine Power System . . . . . . . . . .
6.4.2 Design of Non-negatively Interactive PSSs Installed in
Example Power System . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7 Multi-machine Power Systems Installed with VSC-Based
Stabilizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
with VSC-Based Stabilizers . . . . . . . . . . . . . . . . . . . . . . . . .
7.1.1 Mathematical Model of a Multi-machine Power System
Installed with a Shunt VSC-Based Stabilizer . . . . . . . . .
Installed with a UPFC-Based Stabilizer . . . . . . . . . . . .
7.2 Design of a Shunt VSC-Based Stabilizer by Localized
Phase Compensation Method to Suppress Inter-area Line Power
Oscillations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.2.1 Localized Small-Signal Model of a VSC-Based Unit in a
Multi-machine Power System . . . . . . . . . . . . . . . . . . .
7.2.2 Design of VSC-Based Stabilizer
by Localized Phase Compensation Method . . . . . . . . . .
7.2.3 Robustness of an ESS-Based Stabilizer
to Variation of Line-Loading Conditions . . . . . . . . . . .
7.3 An Example of Multi-machine Power System
with a Grid-Connected FC Power Plan . . . . . . . . . . . . . . . . . .
with a Grid-Connected FC Power Plant . . . . . . . . . . . .
7.3.2 Design of a Stabilizer Attached to the VSC
of FC Power Plant by Localized Phase Compensation
Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . 266
. . 266
. . 272
. . 279
.
.
.
.
.
.
.
.
.
.
282
289
289
301
320
. . 320
. . 324
. . 328
. . 329
. . 329
. . 329
. . 339
. . 349
. . 350
. . 358
. . 362
. . 367
. . 367
. . 375
Contents
7.4 Damping of Multi-mode Oscillations by Multiple

Stabilizers Attached to a Single UPFC . . . . . . . . . .
7.4.1 Coordinated Design of Multiple Stabilizers
by Articial Fish Swarm Algorithm (AFSA) .
7.4.2 Examples . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xi
. . . . . . . . . . 380
. . . . . . . . . . 381
. . . . . . . . . . 383
. . . . . . . . . . 391
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
Chapter 1
Introduction
1.1
Power System Low-frequency Oscillations
Power system low-frequency oscillations are the oscillations of active power

delivered along particular transmission corridors in a power system with the
oscillation frequency from 0.1 Hz up to a couple of Hz. Once started, the oscillations can continue for a while and then disappear, or grow continuously to cause
power system collapse. In this book, they are called power system oscillations or
power oscillations for short.
Manifestation of a power oscillation is the oscillation of relative movement of
angular positions of generators in the power system. The oscillation can be triggered
by severe faults, such as a three-phase to-earth short circuit along or tripping of a
transmission line. It can also occur under normal operating conditions when the
power system is only subject to small disturbances. Hence if the power system
collapse is caused by the power oscillation, it could belong to the problem of power
system large-signal rotor angle (angular) stability or small-signal angular stability [1].
This book is devoted to the introduction of the linearized (small-signal) methods
for the analysis and damping control of the power oscillation. Hence strictly
speaking, the power oscillation discussed in the book should belong to the problem
of power system small-signal angular stability. If the power oscillation exhibits
increasing magnitude, the power oscillation is said to have negative damping and
obviously the power system is unstable in terms of power system small-signal
angular stability. If the power oscillation shows a sustained constant magnitude, or
continues for a certain period of time (over several to tens of seconds) and sets
down eventually, it is reckoned that the oscillation is of poor damping though the
power system is still stable. If the oscillation subsides quickly within several seconds with a damping ratio greater than 0.1 (sometimes it could be 0.05 for
large-scale power systems), it is considered that the power oscillation is well
damped.

H. Wang and W. Du, Analysis and Damping Control of Power System
Low-frequency Oscillations, Power Electronics and Power Systems,
DOI 10.1007/978-1-4899-7696-3_1
Introduction
Power system oscillation was rst observed in the Northern American power
network in Oct. 1964 during a trial interconnection of the Northwest Power Pool
and the Southwest Power Pool [2]. The power oscillation occurred on the tie line
and was of a frequency of 0.1 Hz. Since then, many incidents of power system
oscillations have been reported in power transmission networks around the world.
Examples are as follows:
1. Between the late 1970s and early 1980s, power oscillations were observed in the
power transmission corridor from Scotland to England in the Great Britain
power network. Operational experience indicated that those oscillations were
related to the relatively high loading level of power transmission lines from
Scotland to England. A series of tests were carried out between 1980 and 1985.
Those tests demonstrated that the oscillations occurred when the power transfer
from Scotland to England reached a certain level and the typical oscillation
frequency was around 0.5 Hz [3].
2. Sustained power oscillations were reported in the Taiwan power network in
1984. Those oscillations happened typically when a large amount of power was
transferred along some particular high-voltage transmission lines. It was found
that reduction of the amount of power delivered along the high-voltage transmission lines can effectively improve the damping of power oscillations.
Installation of the PSSs (power system stabilizers) at selected locations worked
successfully to suppress the power oscillations. Further investigation indicated
that other factors, such as the gain value of automatic voltage regulator
(AVR) and characteristics of load, also affected the damping of the power
oscillations [4].
3. An outage of the Northern American Western Systems Coordinating Council
(WSCC) network on August 10 1996 was directly due to the power oscillations.
The power oscillations (with frequency between 0.2 and 0.3 Hz) were triggered
by the loss of a 500-kV line when the system operated under depressed conditions (with overloaded lines and at low voltage level). The incident spread
quickly, tripping more lines, and generating units, which eventually led to the
separation of the WSCC network into four islands. The outage affected
7.5 million customers for up to 9 h, causing considerable economic loss [5].
Over the last half-century, many power system researchers and engineers have
worked on and contributed to the understanding and solution of the problem of
power system oscillations. It is now well recognized that the main cause of power
system oscillations is the poor damping of the so-called electromechanical oscillation modes of the power system. Poor damping could be brought about by the
(1) large amount of long-distance power transmission, (2) weak interconnection of
large power sub-networks, and/or (3) negative damping due to the fast-acting
high-gain AVRs. Power system low-frequency oscillations can be classied,
according to the electromechanical oscillation modes of the power system, into two
types: (1) local power oscillations (associated with local oscillation modes) and
(2) inter-area power oscillations (associated with inter-area oscillation modes).
1.1 Power System Low-frequency Oscillations
Local power oscillations associated with the local oscillation modes normally are
the oscillations of one or a group of local generators against a large power network.
Local generators send power over a long distance to the large power network.
Frequency of the oscillations often is about one or several Hz. Inter-area power
oscillations related to the inter-area oscillation modes are the oscillations between
two or more sub-networks in a large-scale power system. A typical inter-area
oscillation is the tie-line power oscillation between two weakly connected areas in
the power system. The inter-area power oscillation could involve many
sub-networks to oscillate against each other (which sometimes is referred to as an
intra-area oscillation). Normally, the oscillation frequency is lower, from 0.1 to 1 Hz.
A power oscillation could be engaged by one oscillation mode only (local or
inter-area oscillation mode). This is the case referred to as the single-mode power
oscillation when there is only one electromechanical oscillation mode in the power
system being poorly or negatively damped. In the case that there are multiple poorly
or negatively damped electromechanical oscillation modes in the power system, the
power oscillation is participated by multiple oscillation modes. The power oscillation in this case is called the multi-mode power oscillation.
1.2
Linearized Methods for the Analysis and Damping

Control of Power System Oscillations
Study of power system oscillations is in order to investigate the phenomena and to

develop prevention measures from the occurrence of the problem. This is to
understand why the oscillations happen and to recommend how they can be
effectively suppressed. There are two main categories of linearized methods
(small-signal methods) for the analysis and damping control of power system
oscillations: the damping torque analysis and modal analysis.
The technique of damping torque analysis (DTA) was well explained in [6] for a
single-machine innite-bus power system to investigate the effect of excitation
control on power system small-signal angular stability. Reference [6] is the milestone contribution by deMello and Concordia to the eld. Proposal of the DTA is
based on the linearized HeffronPhillips model of the single-machine innite-bus
power system [7, 8]. It is used to examine the electric torque contribution from a
particular source in the power system to the so-called electromechanical oscillation
loop of a generator, which is in fact the linearized swing equation of rotor motion of
the generator. The electric torque is decomposed into two components, synchronizing torque and damping torque. Synchronizing torque is in phase with rotor
angle deviation and damping torque in phase with rotor speed deviation of the
generator. Synchronizing torque is the force to keep the generator in synchronism
with the rest of the power system. Lack of damping torque or contribution of
negative damping torque may lead to poor damping of power system oscillations or
even power system instability.
Introduction
The DTA is established on the physical concept of electric torque and classic
control theory with the description of the system in frequency and Laplace s
domain. It is easy to be understood and simple to be applied. Often analytical
conclusions can be obtained under certain reasonable assumptions. In particular
later, an effective method, the phase compensation method, for the design of a
power system stabilizer (PSS) was developed on the basis of the DTA [911],
which has made the DTA a well-accepted method for the analysis and damping
control of power system oscillations. Since the 1970s, considerable effort has been
spent to extend the DTA for the application in general multi-machine power
systems.
Modal analysis (MA) is based on the modal control theory, a special branch in
modern control theory. It is established on the state-space representation of a power
system, i.e. so-called ABCD state-space model [12]. Applications of modal analysis
in the study of power system oscillations include the following: (1) computation of
eigenvalues and associated eigenvectors (so-called eigensolution) of the state
matrix of the system; (2) computation of controllability index, observability index,
and their product, residue. There is no much difference in the procedure to apply the
MA in a single-machine innite-bus and a multi-machine power system.
Eigensolution is a very specic topic in mathematics. Examples of applying
eigensolution in studying power system oscillations are computation of power
system oscillation modes (i.e. computation of eigenvalues and identication of
oscillation modes among all eigenvalues), eigenvalue sensitivity, and participation
factor [13, 14]. For a large-scale power system, dimension of the state matrix could
become very high. This could lead to the extremely high computational complexity
and even numerical difculty to calculate the eigenvalues of the state matrix. Hence,
it has been a special research topic to compute eigenvalues of interests of a
high-dimensional state matrix considering the features of the large-scale power
system.
Computation of controllability index, observability index, and residue is mainly
applied for the selection of installing locations and feedback signals of stabilizers in
the multi-machine power system. Because computation of controllability index,
observability index, and residue usually needs to perform eigensolution, various
methods to reduce computational cost have been proposed in the literature.
Examples are the selective modal analysis [1315], partial modal analysis [16], and
eigensolution-free modal analysis [17].
The MA is a method based on the results of numerical computation. It can tell
whether a power system is stable at given system operating conditions. However, it
is difcult to be used to draw general conclusions by performing numerical
eigensolution. Hence in many occasions, effort has to be spent to establish the
connections between the computational results from the MA and the physical
explanations in order to understand and explain the results of numerical
computation.
1.3 FACTS and Grid-Connected ESS
1.3
FACTS and Grid-Connected ESS
Flexible AC transmission systems (FACTS) are the technologies emerged around

the end of last century to reinforce power systems. Foundation of the FACTS is the
application of power-electronic based control in high-voltage power transmission
systems. It is the result of rapid advance in high-power electronics technology over
the last several decades. Though the FACTS controllers are still considered
expensive compared to the conventional power system controllers, they have been
installed in many real power systems in the world due to their superior control
performance.
The basic operational principle of the FACTS controllers is shown in Fig. 1.1
[18, 19]. The active power delivered along the transmission line is as follows:
P12
V1 V2
sin h
x
1:1
where h is the phase difference between the voltage at the sending and receiving end
1 and V
2 . A FACTS controller is to change in real time
of the transmission line, V
1 and V
2 indithe line impedance x, and phase and magnitude of line voltage V
vidually or simultaneously to regulate the power flow along the transmission line,
such that a great flexibility of power flow regulation is achieved.
In fact, the operational principle of the FACTS was known as early as in the 1920s.
Since then, mechanically controlled series and shunt capacitors have been installed in
power systems to control line power flow and regulate system voltage prole. Power
electronics applications in high-voltage power systems were dated from early 1970s
when converters for high-voltage direct-current systems began using thyristors to
replace mercury arc valves. This was the time when the FACTS technology really
emerged and was applied in power systems. The concept of the FACTS was formerly
introduced by Prof. Hingarani [18] when work to use power electronic switching
devices to replace thyristors began [19]. Hence normally, the FACTS controllers are
classied into two groups: (1) the conventional thyristor-based FACTS controllers
and (2) new generation of converter (power electronic switching devices)-based
FACTS controllers. The following are three main types of thyristor-based FACTS
controllers which have been applied in power systems:
Fig. 1.1 Illustration on the

operational principle of
FACTS controllers
V1
P12
V2
Introduction
1. Thyristor-controlled static VAR compensator (SVC), which controls the magnitude of line voltage;
2. Thyristor-controlled series compensator (TCSC), which changes the equivalent
impedance of transmission line and hence can be used for power flow control;
3. Thyristor-controlled phase shifter (TCPS), which shifts the phase of line voltage
and normally can be used for power flow control.
New generation of FACTS controllers employs self-commutated, voltage-soured
switching power converters to realize rapidly controllable, static synchronous AC
voltage or current sources. A new-generation FACTS controller is constructed
mainly on the synchronous voltage source (SVS) which is an ideal machine with no
rotation and inertia. The magnitude and phase of the SVS can be controlled
instantaneously to generate reactive power (both capacitive and inductive) and/or to
regulate active power flow independent of power system variables (voltage and
current). For example, Fig. 1.2 shows an SVS constructed by a GTO-based voltage
source converter (VSC), which employs the algorithm of the pulse width modulation (PWM). Through controlling the modulation ratio m and phase /, the
magnitude and phase of the AC SVS can be controlled to realize various functions
of power system regulation.
There are three main types of VSC-based FACTS controllers:
1. The static synchronous compensator (STATCOM). This is the controller realized by installing the VSC-based SVS in shunt along a transmission line.
2. The static synchronous series compensator (SSSC). This is the VSC-based SVS
installed in series in the transmission line.
3. The unied power flow controller (UPFC). This is the most powerful FACTS
controller, a combined device of a STATCOM and a SSSC with a common DC
capacitor.
The development of power systems in the recent years is towards meeting the
requirement of clean power generation. This will lead to the fundamental changes
Three-phase diagram of GTO-based

voltage source converter
Single-line diagram of GTO-based

voltage source converter
Vc
Vdc
transformer
Variable AC
voltage
Power switching A DC
circuit
capacitor
Fig. 1.2 An SVS realized by a VSC circuit
SVS
DC
voltage
m
1.3 FACTS and Grid-Connected ESS
Step-down transformer
Vc
Vc
ac/dc VSC
Cdc
2
nd
converter
ESS
Single stage circuit
ESS
Two stage circuit
Fig. 1.3 VSC-based power electronic circuit for grid-connected ESS or renewable generation
plant
of the systems in the near future with a fairly large percentage of grid-connected
renewable power generation. It can be foreseen that the high-power electronic VSC
will play even more important role in shaping the future power systems. Not only
the VSC will be used for the grid connection of many types of renewable power
generation, such as wind, solar, and fuel cell, but also it can be applied for the
energy storage systems (ESSs) to assist power system operation and control in
accommodating variable grid-connected renewable power generation. Hence in
future power systems, the VSC could become one of the mostly installed power
components like transformers to meet various purposes of power regulation, storage, and generation.
Figure 1.3 shows the conguration of grid connection of an ESS device or a
renewable power generation plant by use of the VSC-based power electronics
circuits. The ESS device can be a battery ESS (BESS). The conguration is also
applicable for the grid connection of a renewable generation source, such as a
photovoltaic (PV) power plant. In the two-stage circuit, the second converter can be
a DC/DC converter to connect to a fuel cell (FC) power plant or a DC/AC converter
to a wind farm. In both circuits, the key component is the AC/DC converter
(VSC) which is connected to the high-voltage busbar via a step-down transformer.
1.4
Controllers to Damp Power System Oscillations
To increase the damping of power system oscillations, installation of a supplementary damping controller, the PSS (power system stabilizer), on the side of
excitation system of a generator is a simple, effective and economical method.
Introduction
Through the effort of and experience accumulated by power engineers and

researchers in the last half-century, the PSS has become a standard device installed
in the generator to supply extra damping to power system oscillations.
However, it has been found that installation of the PSSs cannot always provide
satisfactory solution to the problem of power system oscillations. Typical example
is the inter-area power oscillations on the tie lines or along a particular
long-distance power transmission corridor in a large-scale power system. If there is
no power plant close to the tie lines and the transmission corridor where power
oscillations are observed, it may need complex coordinated design of multiple PSSs
installed at different remote power plants to jointly suppress the power oscillations.
There is no guarantee that the coordinated design of those multiple PSSs is always
able to supply sufcient damping to the power oscillations. Hence, any new option
to the PSS for the effective suppression of power system oscillations is always
welcome.
With the advent of FACTS technology, FACTS-based stabilizers (referred to as
FACTS stabilizers in this book) have been investigated as an alternative type of
controllers to damp power system oscillations. Main difference of the FACTS
stabilizers to PSSs is that the FACTS devices, which the FACTS stabilizers are
attached to, usually are installed at key locations of transmission network in power
systems, such as that close to the tie lines and long-distance transmission corridors.
Hence in certain cases, FACTS stabilizers could be simpler, easier to be implemented, and more effective to suppress power system oscillations.
In fact, the idea of applying variable series compensation to damp power system
oscillations was proposed many decades ago [20, 21], though at that time, variable
compensation was assumed to be realized by switching in or out the series
capacitors using mechanical circuit breakers. Only with the thyristor-controlled
high-power electronic circuits being available after the 1980s, high-speed variable
reactive power compensation has become a feasible technology in power system
application to suppress power system oscillations. Between the late 1980s and
1990s, case study of numerical calculation, computer simulation, and eld tests
conrmed the capability of thyristor-based FACTS stabilizers to damp power
system oscillations. Theoretical analysis was carried out to obtain insight into and to
understand the basic principles of thyristor-based FACTS stabilizers. Similarity of
the thyristor-based FACTS stabilizers to the PSSs has helped considerably the
understanding on why and how the thyristor-based FACTS stabilizers can suppress
power system oscillations.
A FACTS stabilizer is superimposed on the normal control function of a FACTS
controller. For example, Fig. 1.4 shows a TCR (thyristor-controlled reactor)-FC
(xed capacitor) type of SVC with a normal function of voltage control. A damping
controller (the SVC stabilizer) is superimposed on the SVC voltage control function. In Fig. 1.4, the variable inductive current Il is controlled by the thyristor ring
circuit. Hence, the variable equivalent reactance of the TCR-FC SVC is as follows
1.4 Controllers to Damp Power System Oscillations
Power system
Vs
Is
Il
Vs
Vsref
voltage
controller
Ic
x svcl
x svcc
firing circuit
/2
ysvc
SVC
stabilizer
Fig. 1.4 An SVC stabilizer superimposed on the SVC voltage control function
s
s
V
V
jxsvc
Is
Il Ic
1:2
s is the voltage measurement at the location where the SVC

where and in Fig. 1.4, V

is installed, Vsref the reference signal of voltage control, jxsvc the equivalent reactance of the SVC, and ysvc the feedback signal of the SVC stabilizer. The SVC
voltage controller is to maintain the magnitude of voltage at the SVC installing
sref by controlling the variable equivalent reactance of the SVC.
location close to V
Attachment of the SVC stabilizer contributes the control of variable equivalent
reactance jxsvc, hence the exchange of reactive power between the SVC and rest of
the power system. Thus, it affects the variations of active power flow along the
transmission line where the SVC locates. If designed properly, the SVC stabilizer
can help to suppress power oscillations. Attachment of the SVC stabilizer to SVC
normal voltage control function is very similar to the arrangement of a PSS being
superimposed on the voltage control loop of an AVR.
It was well reckoned that VSC-based FACTS stabilizers work as effectively as
the thyristor-based FACTS stabilizers in damping power system oscillations, as
conrmed by case study of numerical calculation, computer simulation, and eld
tests [22, 23]. There is no fundamental difference between the mechanism of the
thyristor-based and VSC-based FACTS stabilizers, though examination of the
VSC-based FACTS stabilizers usually is sometimes more complicated as it
involves the VSC functions.
10
Introduction
Occurrence of power system oscillations is essentially due to the response of

generators to system disturbances in order to maintain the balance of supply and
consumption of active power. Because generators need time to either accelerate or
decelerate in responding to disturbances and they are at different geographic
positions in the power system, power oscillations occur as the result of the relatively
slow correcting actions of generators to keep the balance of active power. The
slower the correcting actions are, the worse the power oscillations. That is why in
the power system, the power oscillations are of low oscillation frequency and hence
called low-frequency oscillations. Gird-connected ESSs do not have rotating
components involved for the exchange of active power with the rest of power
system. They can respond instantaneously to system disturbances with correcting
actions to maintain the balance of active power in the power system. Hence in
principle, applications of the ESSs in the power system can effectively provide extra
damping to power system oscillations.
Research on the capability of ESS controllers to damp power system oscillations
has been mainly about two types of energy storage devices, the superconducting
magnetic energy storage (SMES) and the battery energy storage systems (BESS)
that are also classied as FACTS devices [24]. The function and effectiveness of a
SMES to damp power system oscillations have been well investigated since several
decades ago. By regulating the active and reactive power exchange of the SMES
with the power system, the SMES can effectively suppress power system oscillations, which have been demonstrated by simulation and laboratory experiment [25,
26].
In [27], a case of eld application of the BESS in improving power oscillation
damping was reported. A 10-MW BESS unit was installed in 1988 at the Chino
substation owned by the Southern California Edison (SCE), USA. Six years later,
an energy source power system stabilizer (ESPSS) was installed by taking the
advantage of very fast response rate of the installed BESS to provide extra damping
to the power system. The ESPSS was designed to modulate the active power output
from the BESS to damp power system oscillations rather than through controlling
voltage or reactive power as a conventional stabilizer does.
1.5
Design of Damping Controllers to Suppress Power

System Oscillations
The prime specication of designing a damping controller (a PSS or a FACTS

stabilizer) is the maximum effectiveness of the stabilizer at the minimum control
cost. Two main issues in the design are as follows: (1) selection of installing
locations and feedback signals and (2) parameters setting of the stabilizer.
In a large-scale multi-machine power system, there are many power stations that
can be or have been equipped with the PSSs. To damp a particular oscillation mode
of interests, it would be impractical if all the PSSs in the power system are set
1.5 Design of Damping Controllers to Suppress Power System Oscillations
11
accordingly. Hence, the selection of installing locations of the PSSs can mean to
choose the locations to install PSSs or to choose the PSSs among those installed to
be set, in order to damp the particular oscillation mode. As far as a FACTS stabilizer is concerned, justication of installing an expensive FACTS device usually
is based on other applications rather than the damping function. Hence, the
selection of installing location of the FACTS stabilizer is carried out among the
installed FACTS devices and/or FACTS normal control functions to attach the
damping function. Even if a location in the power system is the best to have the
FACTS stabilizer, it often needs further justication to consider the installation of
the FACTS stabilizer there.
Leaving the practical consideration about the candidate locations aside, there is
no difference in the strategy to select installing locations for the PSSs and FACTS
stabilizers. On the other hand, selection of feedback signals for the PSSs and
FACTS stabilizers normally is to determine the most effective and locally available
feedback signals for the design of stabilizers. Those signals can be the rotor speed
deviation of generators, active power, line current, etc. In the case that remote
signals can be used, of course, it also includes the choice of signals at different
locations in the power system.
Methods which have been proposed and developed so far for the selection of
installing locations and feedback signals are mostly based on the DTA and modal
analysis (MA). Examples are the induced damping and synchronizing torque
coefcients (IDSTCs) index [28], damping torque coefcient (DTC) index [29],
simple index [30], and damping index [31] derived by using the DTA. Sensitivity
calculation [32], the participation factors [1315], the LIED (location index for
effective damping) [33], residue index [34, 35], the partial modal analysis [16], and
eigensolution-free modal analysis [17] are established by use of the modal analysis.
Methods to set parameters of multiple stabilizers in a multi-machine power
system can be classied into two categories, sequence setting and simultaneous
tuning. Sequence setting is the strategy to set stabilizers one by one in a sequence.
For example, suppose two stabilizers, stabilizers A and B, to be set. In sequence
setting, stabilizer A is installed in the multi-machine power system rstly and its
parameters are set. Afterwards, stabilizer B is installed with its parameters being set.
The strategy of sequence setting ts the practice of installing stabilizers in the
multi-machine power system. It is simple and easy to be used. The well-known
phase compensation method [911] is considered to be an approach of sequential
setting.
However, it was found that this strategy of sequential setting can cause a
problem of so-called eigenvalue drifting [36] due to the dynamic interactions of
multiple stabilizers.Eigenvalue drifting is that, for example, setting of stabilizer B
affects the setting results of stabilizer A which is designed previously. The consequence is that the sequential setting cannot complete the required setting of two
stabilizers, A and B, to move oscillation modes to their required positions simultaneously. This phenomenon of eigenvalue drifting is also referred to as the
problem of dynamic interactions between multiple stabilizers.
12
Introduction
In order to solve the problem of eigenvalue drifting caused by interactions

between multiple stabilizers, the simultaneous tuning strategy has been proposed
and developed for the design of multiple stabilizers. In the simultaneous tuning,
parameters of all the stabilizers are tuned simultaneously. Hence, once the tuning is
completed, all the oscillation modes involved in the tuning are at the target positions. Simultaneous tuning is often called the coordinated design of multiple stabilizers. So far there have been many methods of simultaneous tuning for different
types of stabilizers proposed and developed. Those methods to simultaneously tune
parameters of multiple stabilizers can be classied into two groups. The rst group
are the analytical methods developed either from the DTA technique or by applying
the modern control theory. The second group are the methods of parameter optimization. They usually convert the simultaneous tuning to the solution of an
optimization problem by establishing an objective function in the space of stabilizers parameters. Then, an optimization algorithm is employed to nd the solution
of the objective function. Normally, the second group of methods heavily relies on
the eigensolution of the power system.
Other two important issues in the design of stabilizers are the robustness of and
interactions between multiple stabilizers. The robustness is about the effectiveness
of stabilizers when power system operating conditions vary. Because the linearized
model of a power system is established at a chosen operating condition for the
design of stabilizers, the effectiveness of stabilizers is ensured only at the operating
point by the design. Hence, the robustness of stabilizers to the variations of power
system operating conditions is an issue which needs special consideration in order
to guarantee the effectiveness of the stabilizers over a certain range of system
operating conditions. The interactions are those among multiple stabilizers themselves and with other control functions on which they are superimposed. The
well-known examples are the eigenvalue drifting and interaction between a PSS
and the AVR it is attached to.
1.6
Organization of the Book
Linearized methods for the analysis and damping control of power system oscillations covered by this book are based on the linearized models of power systems.
Hence, a comprehensive introduction on establishing linearized models of power
systems is presented in the book. This includes the introduction of the Heffron
Phillips model, a well-known and relatively simpler linearized model of power
system for the study of power system oscillations. In addition, the establishment of
a more general linearized model of power systems is introduced where the full-scale
mathematical model of synchronous generators is used.
Two groups of linearized methods for the analysis and damping control of power
system oscillations are introduced in the book. They are the damping torque
1.6 Organization of the Book
13
Table 1.1 Chapter arrangement of the book

Type of stabilizers
Type of power systems
Single-machine innite-bus
power systems
Multi-machine power
systems
The PSSs
The thyristor-based FACTS
stabilizers
The VSC-based FACTS
stabilizers
Chapter 2
Chapter 3
Chapter 5
Chapter 6
Chapter 4
Chapter 7
analysis (DTA) and the modal analysis (MA). The basic concepts and theory of the
DTA and the MA are presented rstly for a simpler case of power systems, a
single-machine innite-bus power system. The introduction of the applications and
extensions of the DTA and the MA in a more complicated case of a multi-machine
power system is then followed in the book.
Power system damping controllers are classied into three groups in the book,
the power system stabilizer (PSS), the thyristor-based FACTS stabilizers, and the
VSC-based stabilizers. Introduction of linearized methods for the analysis and
damping control of power system oscillations in the book is organized according to
the classication of damping controllers (stabilizers) and their applications in the
simple single-machine innite-bus and complex multi-machine power system. The
organization is illustrated in Table 1.1.
References
1. IEEE/CIGRE Joint Task Force on Stability Terms and Denitions (2004) Denition and
classication of power system stability. IEEE Trans Power Syst 19(2):13871401
2. Schleif FR, White JH (1966) Damping for the northwest-southwest tieline oscillationsan
analogue study. IEEE Trans Power Appar Syst 85(12):12391247
3. Gibson CM (1988) Application of power system stabilizers on the Anglo-Scottish
interconnectionprogramme of system proving tests and operational experience. IEE Proc
Part C 135(3):255260
4. Hsu Y-Y, Shyue S-W, Su C-C (1987) Low frequency oscillations in longitudinal power
systems: experience with dynamic stability of Taiwan power system. IEEE Trans Power Syst 2
(1):9298
5. Western Electricity Coordinating Council (1996) Western systems coordinating council
disturbance report. www.bpa.biz
6. deMello FP, Concordia C (1969) Concepts of synchronous machine stability as affected by
excitation control. IEEE Trans Power Appar Syst 88(4):316329
7. Heffron WG, Phillips RA (1952) Effect of modern amplidyne voltage regulators on
underexcited operation of large turbine generators. AIEE Trans (Power Appar Syst) 71:
692697
8. deMello FP, Laskowski TF (1975) Concepts of power system dynamic stability. IEEE Trans
Power Appar Syst 94(3):827833
9. Larsen EV, Swann DA (1981) Applying power system stabilizers part I: general concepts.
IEEE Trans Power Appar Syst 100(6):30173024
14
Introduction
10. Larsen EV, Swann DA (1981) Applying power system stabilizers part II: performance
objectives and tuning concepts. IEEE Trans Power Appar Syst 100(6):30253033
11. Larsen EV, Swann DA (1981) Applying power system stabilizers part III: practical
considerations. IEEE Trans Power Appar Syst 100(6):30343046
12. Rogers G (2000) Power system oscillations. Kluwer Academic Publisher
13. Perez-Arriaga IJ, Verghese GC, Schweppe FC (1982) Selective modal analysis with
applications to electric power systems, Part I. IEEE Trans Power Appar Syst 101(9):31173125
14. Perez-Arriaga IJ, Verghese GC, Schweppe FC (1982) Selective modal analysis with
applications to electric power systems, part II. IEEE Trans Power Appar Syst 101(9):31263134
15. Sancha JL, Perez-Arriaga IJ (1988) Selective modal analysis of electric power system
oscillatory instability. IEEE Trans Power Syst 3(2):429438
16. Larsen EV, Sanchez-Gasca JJ, Chow JH (1995) Concept for design of FACTS controllers to
damp power swings. IEEE Trans Power Syst 10(2):948956
17. Wang HF (1999) Selection of robust installing locations and feedback signals of
FACTS-based stabilizers in multi-machine power systems. IEEE Trans Power Syst 14
(2):569574
18. Hingorani NG (1988, Aug) High power electronics and flexible ac transmission systems. IEEE
Power Eng Rev 34
19. Gyugyi L (1992) Unied power-flow control; concept for flexible AC transmission systems.
IEE Proc Part C 139(4):323331
20. Smith OJM (1969) Power system transient control by capacitor switching. IEEE Trans Power
Appar Syst 88(1):2835
21. Webster RH, Mane AP, Smith OJM (1971) Series capacitor switching to quench
electromechanical transient in power systems. IEEE Trans Power Appar Syst 99(2):427433
22. Song YH, Johns AT (1999) Flexible AC transmission systems. IEE Press
23. Hingorani NG, Gyugyi L (1999) Understanding FACTS. IEEE Press
24. CIGRE TF 30-01-08 Report (1999) Modelling of power electronics equipment (FACTS) in
load flow and stability programs
25. Hsu CS, Lee WJ (1993) Superconducting magnetic energy storage for power system
applications. IEEE Trans Ind Appl 29(5):990996
26. Xue XD, Cheng KWE, Sutanto D (2005) Power system applications of superconducting
magnetic energy storage systems. In: Fourtieth IAS annual conference meeting on industry
applications, vol 2, Oct 2005, pp 15241529
27. Bhargava B, Dishaw G (1998) Application of an energy source power system stabilizer on the
10 MW battery energy storage system at Chino substation. IEEE Trans Power Syst 13(1):
145151
28. Pourbeik P, Gibbard MJ (1996) Damping and synchronizing torques induced on generators by
FACTS stabilizers in multimachine power systems. IEEE Trans Power Syst 11(4):19201925
29. Swift FJ, Wang HF (1995) Static Var compensator to damp system oscillations in
multi-machine power systems. In: Proceedings of IEE APSCOM conference, Hong Kong
30. Swift FJ, Wang HF (1996) Application of the controllable series compensator in damping
power system oscillations. IEE Proc Part C 143(4):359364
31. Wang HF, Swift FJ (1997) The indexes for selecting the best locations of PSS or
FACTS-based stabilizers in multi-machine power systems: a comparison study. IEE Proc
Part C 144(2):155159
32. Abdalla OH, Hassan SA, Tweig NT (1984) Co-ordinated stabilisation of a multimachine
power system. IEEE Trans Power Appar Syst 103(3):483494
33. Okamoto H, Kurita A, Sekine Y (1995) A method for identication of effective locations of
variable impedance apparatus on enhancement of steady-state stability in large-scale power
systems. IEEE Trans Power Syst 10(3):14011407
References
15
34. Martins N, Lima LTG (1990) Determination of suitable locations for power system stabilizers
and static VAR compensators for damping electromechanical oscillations in large power
35. Yang X, Feliachi A (1994) Stabilization of inter-area oscillation modes through excitation
36. Gooi HB, Hill EF, Mobarak MA, Thonne DH, Lee TH (1981) Coordinated multimachine
stabilizer setting without eigenvalue drift. IEEE Trans Power Appar Syst 100(12):38793887
Chapter 2
A Single-Machine Innite-Bus Power

System Installed with a Power System
Stabilizer
2.1
Linearized Model of a Single-Machine Innite-Bus

Power System Installed with a Power System
Stabilizer
2.1.1
General Linearized Mathematical Model
2.1.1.1
Full Mathematical Model of a Synchronous Generator
Fundamental equations describing the dynamics of a synchronous generator are

the well-known Parks voltage equations. They are given based on a coordinate
system consisting of a d-axis (direct axis) xed on the eld winding axis of synchronous generator and a q-axis (quadrature axis). After Parks transformation,
three armature phase windings a, b, and c on the stator of synchronous generator are
replaced by two equivalent armature phase windings, d and q. Two damper
windings on the rotor, D and Q, are permanently short-circuited. Field winding f is
DC excited. Parks voltage equations of those ve windings have the simplest form
as follows:
w_ d x0 vtd ra id xwq
w_ q x0 vtq ra iq xwd
w_ x0 vf rf if
f
2:1
w_ D x0 rD iD
w_ Q x0 rQ iQ
where wd ; wq ; wf ; wD ; wQ , id ; iq ; if ; iD ; iQ , vtd ; vtq ; vf , and ra ; rf ; rD ; rQ are the magnetic flux linkage, current, voltage, and resistance of each corresponding winding,
respectively, x0 is the synchronous speed, and x is the rotor speed in per unit (p.u.)
DOI 10.1007/978-1-4899-7696-3_2
17
2 A Single-Machine Innite-Bus Power System
18
of synchronous generator. Since there is no static coupling between any d-axis

winding and q-axis winding, the relationship between the magnetic flux linkage and
the current of those windings is as follows:
2
wd
xd
xad
xad
32
id
6
7 6
76
7
4 wf 5 4 xad xf xad 54 if 5
wD
xad xad xD
iD
"
#

wq
xq xaq iq
wQ
xaq xQ
iQ
2:2
where xd ; xq ; xf ; xD ; xQ are the self-reactance of winding d, q, f, D, and Q,

respectively. In Eq. (2.2), it is assumed that the mutual reactance of all windings on
a common axis is same, being xad or xaq respectively.
The rotor motion equation of synchronous generator is as follows:
:
d x0 x 1
1
:
x Tm Tt Dx 1
M
2:3
where M is the inertia of the rotor, D the damping coefcient of the rotor motion, d
the rotor angular position of synchronous generator to a reference axis, and Tm and
Tt the mechanical torque and electric torque applied on the rotor of generator,
respectively.
2.1.1.2
Excitation System and the Automatic Voltage Regulator

(AVR)
Voltage control in a power system is closely related to the regulation of reactive

power flow. The objectives of voltage and reactive power control in the power
system are as follows:
1. To maintain the voltage at terminals of all equipment in the power system within
acceptable limits. As far as the system voltage prole is concerned, a constraint
of system voltage deviations is imposed normally to be within 5 % of the
nominal value of voltage.
2. To improve system stability in order to maximize the utilization of the transmission system.
3. To minimize reactive power flow so as to reduce transmission losses.
Power system voltage control has a hierarchy structure with multiple levels. At
the primary level, control devices attempt to compensate the rapid and random
voltage variations by maintaining their output variables close to the setting reference values. The highest level of voltage control uses global information of power
2.1 Linearized Model of a Single-Machine

Fig. 2.1 Arrangement of an
AVR
19
if
vf0
vf
vf'
Vt
TE(s)
Vtref
+
+ u pss
AVR
system and is implemented by solving some optimization problems. A synchronous

generator can generate or absorb reactive power depending on its excitation, which
is controlled by a voltage control device, the automatic voltage regulator (AVR).
Figure 2.1 shows simple illustration on the arrangement of the AVR on the synchronous generator.
In Fig. 2.1, TE(s) denotes the transfer function of the AVR. The excitation
provided to the eld winding of synchronous generator is vf which is from two
sources. One is a constant excitation, vf0 , and another is the output from the AVR,
v0f . The AVR measures the terminal voltage of synchronous generator, Vt , and
compares it to a reference setting value Vtref . The AVR responds to the deviation of
terminal voltage to change the excitation of generator and hence regulates the
reactive power supply or absorption by generator. By doing so, the terminal voltage
of generator is effectively maintained close to the reference setting value. Therefore,
the AVR works at the bottom of the hierarchy of voltage control of power system to
eliminate voltage variations at generators terminal. On the other hand, the reference setting value of the AVR is provided by the result of voltage and/or reactive
control optimization of the whole power system such that the voltage prole of the
system is kept within required constraints and the transmission losses are minimized. Hence, Vtref is from the higher level of hierarchy of voltage control of the
power system.
Historically, the role of excitation system to improve power system performance
has been growing. Early excitation systems were for the maintenance of constant
voltage level at generators terminals and were implemented manually. Gradually,
fast-acting AVRs were installed in many generators not only providing a satisfactory voltage control performance, but also considerably improving power system
steady-state and transient stability. However, during 1960s to 1970s, it was found
that fast-acting AVRs have an adverse effect in providing negative damping to
power system oscillations in some occasions. This results in a conflict requirement
in the design of the AVRs. Subsequently, the power system stabilizer (PSS) was
introduced into excitation control to overcome the problem. Nowadays, it is quite
common to have a combination of a fast-acting AVR and a PSS for the excitation
control of synchronous generators. Stabilizing signal, upss , is superimposed on that
of the AVR as shown in Fig. 2.1.
20
synchronous generator
DC motor
slip ring
field winding of
dc motor
Vt
Vtref
AVR
+
Fig. 2.2 Arrangement of a DC excitation system
Based on the difference of the excitation power sources used, excitation systems
can be classied into three major types.
1. DC excitation systems
A DC excitation system uses a DC generator as the source of excitation power to
provide eld current and is connected to eld winding through slip rings. The
exciter may be driven by a motor or by the generator itself. The DC excitation
system represents the early application of excitation control. Figure 2.2 shows the
arrangement of the DC excitation system.
2. AC excitation systems
An AC excitation system uses an AC machine (alternator) as the source of excitation power to provide eld current. The AC excitation current is rectied to
provide the DC excitation to the synchronous generator. Usually, the AC exciter is
on the same shaft of synchronous generator. There are two major types of AC
excitation systems, depending on the difference of the arrangement of AC excitation
from either the stationary or rotating armature winding of the exciter, as shown in
Figs. 2.3 and 2.4, respectively. In a rotating rectier excitation system, the armature
windings of the AC exciter and the diode rectiers rotate with the synchronous
generator eld. Thus, the need for slip rings and brushes is eliminated. Hence, such
a system is also called a brushless excitation system.
3. Static excitation systems
In a static excitation system, power supply is from the synchronous generator.
The DC excitation is provided to the eld of generator through slip rings after being
rectied. There are three major types of static excitation systems. They are
potential-source controlled-rectier systems, compound-source rectier systems,
and compound-controlled rectied excitation systems. Figure 2.5 shows the
arrangement of a potential-source controlled-rectier excitation system.

alternator
21
stationary diode
slip ring
Vt
field winding of
alternator
Vtref
AVR
+
Fig. 2.3 Arrangement of a stationary rectier excitation system
alternator
rotational armature and

diode rectifier
field winding of
alternator
Vt
Vtref
AVR
+
Fig. 2.4 Arrangement of a rotating rectier excitation system
controlled rectifier
AVR
Vt
Vtref
Fig. 2.5 Arrangement of a potential-source controlled-rectier excitation system
22
From Fig. 2.1, it can have

vf vf0 TEsVtref Vt upss
2:4
where vf0 is the constant excitation, upss the stabilizing signal of the PSS, and Vt
and Vtref the terminal voltage of generator and its reference setting value, respectively. Various forms of transfer function have been recommended for different
types of excitation systems and the AVR. In this book, the following simplest form
is used for the purpose of simple presentation,
TEs
KA
1 sTA
2:5
where KA is the gain and TA the time constant of the AVR. Hence, from Eqs. (2.4)
and (2.5), mathematical model of the AVR can be written as
vf vf0 v0f

1
KA
Vtref Vt upss
v_ 0f v0f
TA
TA
2.1.1.3
2:6
A Single-Machine Innite-Bus Power System
Figure 2.6 shows the conguration of a power system where a generator sends
power to a large network. Capacity of the large network is much greater than that of
the generator such that operation of the large network is not affected at all by any
changes in the part of the power system on the left-hand side of busbar b in Fig. 2.6.
This effectively means that the voltage and frequency at busbar b are constant when
the focus of the study is the part of the left-hand side of the power system. Thus,
from the point of view of operation of the part of left-hand side of the power
system, capacity of the large network is innite. Hence, busbar b is called the
innite busbar, and the part of the power system on the left-hand side of busbar b
is a single-machine innite-bus power system. The single-machine innite-bus
power system is an approximate representation of a kind of real power systems,
where a power plant with a generator or a group of generators are connected by
transmission lines to a very large power network.
Vt
Vb
xt
A large network
Pt
It
busbar b
Fig. 2.6 A single-machine innite-bus power system
23
Fig. 2.7 dq coordinate of

generator
Vt
jxt I t
Vb
vq
It
iq
id
vd
For the single-machine innite-bus power system shown in Fig. 2.6, it can have
t jxtIt V
b
V
2:7
In the dq coordinate of the generator shown in Fig. 2.7, Eq. (2.7) can be written
as follows:
vtd jvtq jxt id jiq vd jvq
2:8
where vtd ; vtq and id ; iq vd ; vq are the d and q components of terminal voltage of
t , line current, It , and voltage at the innite busbar, V
b , respectively.
generator, V
Comparing the real and imaginary part on the both sides of Eq. (2.8), it can have
vtd xt iq vd
vtq xt id vq
2:9
where vd Vb sin d; vq Vb cos d and the terminal voltage of generator is as

follows:
Vt
q
v2td v2tq
2:10
In per unit, the mechanical and electric torque, Tm and Tt , in Eq. (2.3), is equal to
the mechanical power input from the prime mover to the electric power supplied by
the generator, respectively, i.e. Tm Pm and Tt Pt . While Pt is equal to the
electric power received at the innite busbar, that is
Pt vdt id vqt iq vd id vq iq
2:11
24
Equations (2.1)(2.3), (2.6) and (2.9)(2.11) are the complete dynamic model of
the single-machine innite-bus power system shown in Fig. 2.6 where Vb and Pm
are constant.
2.1.1.4
Linearized Model of Single-Machine Innite-Bus Power

System
Linearization of Eqs. (2.1) and (2.2) is as follows:

Dw_ d x0 Dvtd ra Did x0 Dwq wq0 Dx
Dw_ q x0 Dvtq ra Diq x0 Dwd wd0 Dx
Dw_ x0 Dvf rf Dif
f
2:12
Dw_ D x0 rD DiD

Dw_ Q x0 rQ DiQ
2
Did
xd
xad
xad
31 2
Dwd
7 6
7 6
7
6
4 Dif 5 4 xad xf xad 5 4 Dwf 5
DiD
xad xad xD
DwD
#

1 "
Dwq
xq xaq
Diq
DwQ
xaq xQ
DiQ
2:13
where prex, D, and subscript 0 are used to denote small increment of a variable
(linearized variable) and value of the variable at the power system steady-state
operating condition where the linearization is carried out, respectively. This notation will be used throughout this book.
Linearization of Eqs. (2.10) and (2.11) is as follows:
DVt
vtq0
vtd0
Dvtd
Dvtq
Vt0
Vt0
DPt vdt0 Did vqt0 Diq id0 Dvdt iq0 Dvqt
2:14
2:15
By using Eqs. (2.14) and (2.15), linearization of Eqs. (2.3) and (2.6) can be
obtained to be
Dd_ x0 Dx
1
Dx_ DPt DDx
M
1
vdt0 Did vqt0 Diq id0 Dvdt iq0 Dvqt DDx
M
2:16
25
Dvf Dv0f
Dv_ 0f

vtq0
1
KA
vtd0
0
Dvf

Dvtd
Dvtq Dupss
TA
TA
Vt0
Vt0
2:17
Arranging Eqs. (2.12), (2.13), (2.16), and (2.17) in matrix form with all linearized current variables be cancelled, it can have
sDXgdq Agdq DXgdq Bgdq DVdq bpss Dupss
DIdq Cgdq DXgdq
2:18
where

DXgdq Dd
Dx
Dvf
T
DVdq Dvtd
Dwd
Dwq
Dvtq ; DIdq Did
Dwf
Diq
DwD
DwQ
For the single-machine innite-bus power system, the network voltage equation
is Eq. (2.9). Its linearization is as follows:
Dvtd xt Diq Dvd xt Diq Vb cos d0 Dd
Dvtq xt Did Dvq xt Did Vb sin d0 Dd
2:19
In matrix form, the above equation can be written as follows:

DVdq Fdq1 DIdq Fdq2 DXgdq
2:20
where

Fdq1
0
xt

xt
Vb cos d0
; Fdq2
Vb sin d0
0
0
0
Substituting Eq. (2.20) into (2.18), state-equation model of the single-machine

innite-bus power system is obtained to be
sDXgdq Agcdq DXgdq bpss Dupss
where Agcdq Agdq Bgdq Fdq1 Cgdq Bgdq Fdq2 :
2:21
26
2.1.2
HeffronPhillips Model
2.1.2.1
Simplication
For the study of power system oscillations, full mathematical model of synchronous
generator of Eqs. (2.1)(2.2) can be simplied based on the following
considerations:
1. Effect of damper windings is not considered or directly included in the damping
coefcient D in the rotor motion equation in Eq. (2.3). Thus, Eq. (2.1) is simplied to be
w_ x0 vtd ra id xw
d
w_ q x0 vtq ra iq xwd
w_ f x0 vf rf if
2:22
2. Effect of fast transient and the resistance of d and q armature windings are
neglected. Equation (2.22) is further simplied to be
0 vtd xwq
0 vtq xwd
2:23
w_ f x0 vf rf if
3. In small-signal power oscillations, variation of rotor speed is very small, x 1.
Hence, the rst two equations in Eq. (2.23) become
vtd wq
vtq wd
2:24
To transform the third equation in Eq. (2.23) into a different form, it is dened
that
E0q
xad
xad vf
w ; Eq xad if ; Efd
xf f
rf
2:25
where E0q is called the q-axis transient excitation voltage, Eq the q-axis excitation
voltage, and Efd the excitation voltage. Multiplying both sides of the third equation
in Eq. (2.23) by xradf , it can have
27
0
T0d0 E_ q Efd Eq
2:26
where T0d0 xx0 rf f , which is the time constant of the eld winding.
Equation (2.26) together with Eq. (2.3) forms the simplied third-order model of
synchronous generator. Equation (2.2) becomes
wd
xd
xad
wf
xad xf
wq xq iq
id
if
2:27
From Eqs. (2.24), (2.25), and (2.27), it can have

vtd wq xq iq
vtq wd xad if xd id Eq xd id
2:28
From Eqs. (2.25) and (2.27), it can be obtained that

E0q

xad
xad
x2
wf
xf if xad id Eq ad id Eq xd x0d id
xf
xf
xf
2:29
x2
where x0d xd xadf , which is called the transient d-axis reactance. Thus, Eq. (2.26)
becomes

0
T0d0 E_ q Efd E0q xd x0d id
2.1.2.2
2:30
A Simplied Model of Single-Machine Innite-Bus Power

System
For the single-machine innite-bus power system shown in Fig. 2.6, from Eqs. (2.9)
and (2.28), it can have
vtd vd xt iq xq iq
vtq vq xt id Eq xd id E0q x0d id
2:31
Thus,
vd xt xq iq xqR iq
vq Eq xd xt id E0q x0d xt id Eq xdR id E0q x0dR id
2:32
28
Fig. 2.8 Equivalent circuit
model of single-machine
innite-bus power system
id
iq
+
x d (x d ')
x q
vq
vd
Eq (Eq ')
d winding
Fig. 2.9 Phasor diagram of

single-machine innite-bus
power system
q winding
EQ
j(x q x 'd )I t
Eq '
jx 'd I t
jx t I t
Vt
vq
Vb
iq
It
id
vd
The single-machine innite-bus power system can be represented by a circuit

model of Fig. 2.8. Figure 2.9 shows the phasor diagram of the system on the dq
coordinate.
From Eq. (2.32) or Fig. 2.8, it can have
E0q Vb cos d
x0dR
Vb sin d
iq
xqR
id
2:33
By substituting Eqs. (2.9) and (2.33) into Eq. (2.11), the electric power supplied
by the generator can be expressed as follows:
29
0
Eq Vb cos d
Vb sin d
Vb sin d
xqR
x0dR

0
Eq Vb
V2 xq x0d
0 sin d b 0
sin 2d
xdR
2 xdR xqR
Pt = Vb cos d
2:34

E0q Vb cos d
Eq E0q xd x0d id E0q xd x0d
x0dR

0
0
Eq xdR
xd xd Vb cos d
0
x0dR
xdR
2:35
where xdR xd xt . From Eqs. (2.31) and (2.33), it can have

Vb sin d xq Vb sin d
;
xqR
xqR
2:36
E0q Vb cos d xt E0q Vb x0d cos d
vtq = Vb cos d xt id Vb cos d xt

x0dR
x0dR
x0dR
vtd = Vb sin d xt iq Vb sin d xt
Hence, the simplied model of single-machine innite-bus power system is as

follows:
:
d x0 x 1
1
:
x Pm Pt Dx 1
M

1
0
E_ q 0 Eq Efd
Tdo

1
KA
0
E_ fd E0fd
Vtref Vt upss
TA
TA
2:37
where

E0q Vb
V2b xq x0d
sin 2d
Pt 0 sin d
xdR
2 x0dR xqR

E0q xdR
xd x0d Vb cos d
Eq 0
x0dR
xdR
Efd
vtd
Efd0 E0fd
q
xt E0q Vb x0d cos d
xq Vb sin d
; vtq 0
; Vt v2td v2tq
0
xdR
xqR
xdR
2:38
30
The model is a group of 4 rst-order differential equations plus 6 algebraic

equations.
2.1.2.3
HeffronPhillips Model [13]
By linearizing Eqs. (2.37) and (2.38) at an operating point of power system, where
Vt Vt0 ; Vtd Vtd0 ; Vtq Vtq0 ; d d0 ; x0 1; E0q E0q0 ; Efd Efd0 , it can have
Dd_ xo Dx
1
Dx_ DPt DDx
M
1
0
DE_ q 0 DEq DE0fd
Tdo
1
KA
0
DE_ fd DE0fd
DVt Dupss
TA
TA
2:39
DPt K1 Dd K2 DE0q
DEq K3 DE0q K4 Dd
DVt
2:40
K5 Dd K6 DE0q
where
K1
E0q0 Vb x0dR
cos
V 2 xq x0
d0 bx0 xqR d cos 2d0
dR
K2 xV0 b sin d0
dR
K3 xxdR
0
dR
xd x0d Vb sin d0
K4
x0dR
X
V V x0 sin d0
q Vb cos d0
td0
K5 V
Vtq0 b0 xd0
xqR
Vt0
t0
dR
V
K6 Vtq0 xx0 t
t0
dR
Substituting Eq. (2.40) into Eq. (2.39), it can be obtained that

Dd_ xo Dx
1
Dx_ K1 Dd K2 DE0q DDx
M
1
0
_
DEq 0 K3 DE0q K4 Dd DE0fd
Tdo
1
KA
0
DE_ fd DE0fd
K5 Dd K6 DE0q Dupss
TA
TA
2:41
31
K1
Pt
Ms + D
K4
K2
Eq
Td0's+K3
K5
_
KA
1+sTA
+ upss
K6
Fig. 2.10 HeffronPhillips model of a single-machine innite-bus power system
Equation (2.41) is the so-called HeffronPhillips model of single-machine

innite-bus power system, which is shown in Fig. 2.10.
The HeffronPhillips model can be written in the form of state-space representation of Eq. (2.21) where
2
DXgdq
bpss
Dd
6 Dx
6
6
4 DE0q
DE0fd
2 3
0
6 7
6 0 7
7
6
6 0 7
4 5
KA
TA
3
7
7
7; Agcdq
5
0
6
1
6 K
M
6
6
6 K0 4
6
Tdo
4
K
TA K5
A
xo
D
M
0
0
0
2
K
M
K0 3
Tdo
K
TA K6
A
7
0 7
7
7
1 7;
T0do 7
5
T1
A
2:42
32
2.2
Modal Analysis
2.2.1
Basis of Modal Analysis Theory
2.2.1.1
Modal Decomposition
State-space representation of a linear system is as follows:

sX Ao X bo u
y cTo X
2:43
u Hsy
where Ao , bo and cTo is the state matrix, control vector, and output vector of
open-loop system, respectively, and Hs is the transfer function of feedback
controller. Transfer function of open-loop system is as follow:
Gs
y
cTo sI Ao 1 bo
u
2:44
The system is shown in Fig. 2.11. Transfer function of closed-loop system is as

follows:
Ts
y
Gs
w 1 GsHs
2:45
Eigen solution is one of the basic techniques in the modal analysis, involving the
computation of eigenvalues and eigenvectors of state matrix, Ao . An eigenvalue of
matrix Ao , k, is a scalar parameter, which satises the following equation
Ao v kv
with a non-trivial solution v 6 0.
Fig. 2.11 Block diagram of a closed-loop control system
2:46
2.2 Modal Analysis
33
Obviously, Eq. (2.46) can be written in the following form

Ao kIv 0
2:47
where I is an unity matrix. In order for Eq. (2.47) to have the non-trivial solution, it
should have
jAo kIj 0
2:48
Equation (2.48) is the following polynomial equation if Ao is an M M matrix

1M kM aM1 kM1 a1 k a0 0
2:49
which is called the characteristic equation of state matrix Ao . The characteristic

equation should have M solutions; that is, matrix Ao has M eigenvalues, if the
dimension of matrix is M.
For the ith eigenvalue of matrix Ao , ki , if a nonzero vector vi satises the
equation
Avi ki vi ;
i 1; 2; . . . ; M
2:50
vi is called the right eigenvector of matrix A associated with ki . Equation (2.50)

can be arranged as follows:
2
A v1
v2
. . . vM v1
v2
k1
60
6
. . . vM 6
40
0
0
k2
0
0
..
.
0
0
3
0
0 7
7
7
0 5
kM
2:51
That is
V1 AV K
2:52
where
2
V v1
v2
...
k1
60
6
vM ; K 6
40
0
0
k2
0
0
0
0
..
.
0
3
0
0 7
7
7
0 5
kM
34
Denote
2
V
1
wT1
6 wT
6 2
6
6 ..
4 .
3
7
7
7 w1
7
5
. . . w M T WT
w2
2:53
wTM
From Eqs. (2.52) and (2.53), obviously it can have
wTi A wTi ki ;
i 1; 2; . . .; M
2:54
Hence, wTi is called the left eigenvector corresponding to eigenvalue ki .

If a new state variable vector z is introduced and dened to be
X VZ
2:55
From Eqs. (2.43), (2.52), and (2.53), it can be obtained that

sZ KZ WT bo u
2:56
y cTo VZ
That is
szi ki zi ; wTi bo u
y cTo
M
X
i 1; 2; . . .; M
2:57
vi z i
i1
According to Eq. (2.57), the system can also be shown in Fig. 2.12. This is the
modal decomposition representation of state-space model of open-loop system.
w1T b 0
1
s 1
z1
w 2Tb0
1
s 2
z2
c0 T v1
c0 T v 2
w M Tb0
1
s M
zM
c0 T v M
Fig. 2.12 Block diagram of modal decomposition representation of state-space model of

open-loop system
2.2 Modal Analysis
2.2.1.2
35
Stability of Open-Loop System and Closed-Loop System
Considering the open-loop system when u 0, the state-space representation of

Eq. (2.57) is as follows:
szi ki zi
2:58
Solution of Eq. (2.58) is as follows:

zi t zi 0eki t ;
i 1; 2; . . .; M
2:59
where zi 0 is the initial value of state variable zi t; i 1; 2; . . .; M. From

Eqs. (2.55) and (2.59), it can have
2
z1 0ek1 t
6 z2 0ek2 t
6
X V6
..
4
.
3
7
7
7
5
2:60
zn 0ekM t
Hence, time response of the

xk t; i 1; 2; . . .; M, is as follows:
kth
state
variable
xk t vk1 z1 0ek1 t vk2 z2 0ek2 t vkM zM 0ekM t
of
M
X
the
system,
vki zi 0eki t
i1
2:61
Obviously, the time response of system state variables is decided by the
eigenvalues of state matrix Ao . If there is one or more eigenvalues on the right-hand
half of the complex plan (the real part of eigenvalue is equal to or greater than zero),
the system is unstable. If all the eigenvalues of Ao are on the left-hand side of the
complex plane, the system is stable. Hence, eigenvalues of Ao determine the system
stability. They often are called the modes of the system. If a pair of eigenvalues are
i;i 1 n jxi , the corresponding component in
conjugate complex number, i.e. k
i
the time response of the kth state variable of the system should be

vki zi 0eki t vki zi 0eni jxi t vki zi 0eni t cos xi t j sin xi t
2:62
The component is oscillatory in respect of time. The oscillation angular frequency is xi . The decaying and increasing of the oscillation are determined by the
36
real part of the mode ni . The pair of conjugate eigenvalues of state matrix Ao are
often called the oscillation mode of the system.
i;i 1
The oscillation frequency f i (Hz) and damping fi associated with k
ni jxi are normally dened as follows:
fi
xi
ni
; f q
2p i
2
n x2
i
2:63
From Eq. (2.59), it can be seen that zi t; i 1; 2; . . .; M is related only with the
ith mode of the system ki . Hence, zi t; i 1; 2; . . .; M often is seen as the ith mode
of the system. Equation (2.57) is often called the modal decomposition of
state-space representation.
From Eq. (2.61), it can also be seen that the magnitude of vki measures how
much the ith mode ki contributes to the kth state variable xk t. Thus, jvki j is a kind
of measurement of the observability of the ith mode in the kth state variable.
On the basis of above discussion, from Fig. 2.12, it can be seen that wTi bo is the
weight on how much the control signal u affects the ith mode of the open-loop
system, the so-called controllability index, whereas cTo vi is the weight on how much
the ith mode is observed in the system output, which is called the observability
index. The product of controllability and observability index is called the residue.
That is
Ri wTi bo cTo vi
2:64
From Eqs. (2.53) and (2.55), it can have

Z V1 X WT X
2:65
zi t w1i x1 t w2i x2 t wMi xM t
2:66
or
wki is the ith row kth column element of matrix W. Equation (2.66) indicates that
the magnitude of wki measures the influence of the kth state variable xk t on the ith
state variable zi t, or the ith mode ki of the system. It is a kind of measurement of
controllability of the kth state variable on the ith mode.
Let the realization of the transfer function of feedback controller Hs be
sXf Af Xf bf y
u cTf Xf
2:67
2.2 Modal Analysis
37
That is Hs cTf sI Af 1 bf . Thus, from Eqs. (2.43) and (2.67), the

state-space representation of closed-loop system can be obtained to be
sX
sXf
Ao
bf cTo
bo cTf
Af
X
Xf

Ac
X
Xf

2:68
where Ac is the state matrix of closed-loop system. Obviously, based on the discussion above, eigenvalues of Ac or modes of closed-loop system determine the
stability of closed-loop system.
From Eq. (2.68), it can be obtained that
Ac Ao Hsbo cTo
2:69
Denote a variable parameter of feedback controller as a. Thus, state matrix is a

function of the parameter. Influence of the parameter on the ith mode of closed-loop
system can be calculated by use of the following equation
@ki
@Ac a
@Hki ; a T T
@Hki ; a
vi
wi bo co vi Ri
wTi
@a
@a
@a
@a
2:70
Hence, the residue measures how much the mode of closed-loop system is
affected by the parameter of the controller.
2.2.2
Applications of Modal Analysis
2.2.2.1
Modal Analysis for the AVR
Consider the simple case of a single-machine innite-bus power system expressed

by
model shown in Fig. 2.10 without the PSS installed
the HeffronPhillips

Dupss 0 . The upper part can be considered as the open-loop system and lower
part the feedback controller. Thus, Eq. (2.41) can be rearranged as follows:
sDd
sDx
"
#
"
#
xo Dd
0
DE0q
2
D
K
M
Dx
M

Dd
0
Dx
0
1
K
M
Dd 1
DE0q Fdelta sDd
2:71
38
In the above state-space model of the power system, the input to the open-loop
system is DE0q , the output is Dd, and the transfer function of feedback controller is
Fdelta s. Obviously, the state-space realization of Fdelta s is as follows:
3
2
3
1

0
K0 3
K0 4
0
DE
Tdo 5
Tdo
Tdo 7
6
q
Dd
4
4
5
0
K
K
1
DE
A K6
fd
T
T
TA K5
A
A
A

0
DEq
DE0q 1 0
DE0fd
sDE0q
sDE0fd
2:72
According to Eq. (2.48), the modes of open-loop system can be found by solving
the following characteristic equation

xo
1 0 k
D k 0 1 K1
M
M

0

K1
M

xo
D
K1
2
D k k M k M xo 0
M
2:73
Solution of Eq. (2.73) gives the oscillation mode of the single-machine

innite-bus power system when DE0q 0, that is the case when the generator is
modelled as a constant voltage source. The oscillation mode is as follows:
1;2
k
2
3
s

14 D
D 2 4x0 K1 5

no jxNOF

2
M
M
M
2:74
The oscillation mode is related to the rotor motion of generator, i.e. state variables Dd and Dx. It is often called the electromechanical oscillation mode of the
power system.
From Eq. (2.71), it can have

0
sDd
K
K
1
M M2 Fdelta s
sDx
xo
D
M

Dd
Dd
Ac
Dx
Dx
where Ac is the state matrix of closed-loop system.
2:75
2.2 Modal Analysis
39

2
Fdelta s 1
31 2
3
K0 4
T 7 4
Tdo 5
5
K
1
TA K5
s T
A
A
2
s K0 3
Tdo
6
0 4
KA K6
TA
0
do
s K0 3
Tdo

1

s T1 KTA K6 10
Tdo
A
A
h
T0do TA
s T1
A
sT0do K3 TA s 1 KA K6
32
3
1
K0 4
T0do
Tdo 5
54
K
s K0 3
TA K5
Tdo
A
2
3
i K0 4
1 4
Tdo 5
T0do
K
TA K5
A
s T1
A
4
0
K
K
TA 6
A
TA s 1K4 KA K5
K3 TA s 1 KA K6
sT0do
2:76
Hence, according to Eq. (2.70), it can have
TA ki 1K4 KA K5
@ 0

k
T
@ ki
i do K3 TA ki 1 KA K6
Ri
@a
@a
2:77
where a KA or TA . By using Eq. (2.77), effect of the AVR on the damping of

electromechanical oscillation, i.e. power oscillation, can be examined.
2.2.2.2
Modal Analysis for the PSS
The general linearized model of the single-machine innite-bus power system with
the PSS installed is Eq. (2.21) which can be rearranged as follows:
2
3 2
0
sDd
6
7 6 a21
4 sDx 5 4
M
sDx38
a138
3 2
3
0
Dd
7 6
7
aT238 76
54 Dx 5 4 0 5Dupss
M
bpss3
Dx38
A33
2
3
Dd
6
7
1 0 4 Dx 5
x0
a22
M
a238
y Dx 0
32
Dx38
Dupss Tpss sDx
2:78
40
a 21
1
Ms + a 22
0
s
Te
a 23-8 T
a 23-8
a13-8
X 3-8
( sI - A 33 ) 1
u pss
bpss-3
Fig. 2.13 General linearized model of single-machine innite-bus power system with PSS
installed
The model is shown in Fig. 2.13.

i and the
Let the oscillation mode of the system without the PSS installed be k
corresponding left and right eigenvector be

wTi wi1
wi2
wTi3
3
vi1
; vi 4 vi2 5
vi3
2:79
That is

wi1
wi2
wTi3
0
4 a21
M
a138
x0
a22
M
a238
aT238
M
A33

i wi1
5k
wi2
wTi3
2:80
It can have
a21
i w
Ti3 a138 k
i1
w
M
a22
i w
Ti3 a238 k
i2
i1 w
i2
w
x0 w
M
aT
i w
Ti3 A33 k
Ti3
i2 238 w
w
M
i2
w
2:81
2.2 Modal Analysis
41
Hence,
Ti3 w
i2
w
aT238
ki I A33 1
M
2:82
From Eqs. (2.64), (2.78), and (2.82), the residue can be obtained to be

Ri wi1
6
Ti3 4
w
wi2
0
0
3
7
5 0 1
bpss3
2.2.2.3
wTi3 bpss3 vi2

aT
i2 vi2 238
w
M
3
vi1
6 7
0 4 vi2 5
vi3
2:83
i I A33 1 bpss3
k
Design of PSS by Pole Assignment
If the feedback signal and transfer function of the PSS to be designed is y and
TPSS s respectively, Eq. (2.78) can be written more generally as follows:
sX AX bDupss
y cT X
Dupss Tpss sy
2:84
Transfer function of open-loop system is as follows:

Gs cT sI A1 b
2:85
Characteristic equation of closed-loop control system is as follows:

1 GsTpss s 0
2:86
If design of the PSS is to assign the electromechanical oscillation mode of the

c n jxc ,
single-machine innite-bus power system to a target position, k
c

kc must be the solution of the characteristic equation of closed-loop control system
of Eq. (2.86). Hence, it should have
c Tpss k
c 0
1 Gk
2:87
42
By separating the real and imaginary part of the above equation, two equations
will be obtained which can be used to determine two parameters of the transfer
function of PSS. If the transfer function of PSS adopts the following format of a
leadlag block,
Tpss s Kpss
1 saT2
2:88
1 sT2
With a predetermined T, parameters of the PSS, Kpss and a , can be set according
to Eq. (2.87), thus completing the design of PSS via the pole assignment.
2.3
Damping Torque Analysis
2.3.1
Damping Torque and Synchronizing Torque
2.3.1.1
Damping Torque and Synchronizing Torque Derived

from HeffronPhillips Model
The damping torque analysis (DTA) was rstly introduced on the basis of the
HeffronPhillips model for a single-machine innite-bus power system to examine
the effect of excitation control, such as the AVR, on power system small-signal
stability [13]. It was developed based on the understanding that the dynamic of the
electromechanical oscillation loop of a generator decides the damping of power
oscillations in the single-machine innite-bus power system.
In the HeffronPhillips model shown in Fig. 2.10, the upper part obviously is the
linearized rotor motion equation and lower part is formed from the mathematical
description of dynamic of the eld winding of generator and the AVR. Figure 2.14
shows the upper part of the model which is called the electromechanical oscillation
loop. Signal DTe from the lower part in the HeffronPhillips model is obviously an
electric torque. Hence, from Fig. 2.14, it can have
K1
Electromechanical oscillation loop
1
Ms + D
0
s
Te
Fig. 2.14 Electromechanical oscillation loop of generator
2.3 Damping Torque Analysis
s2 Dd
43
D
x0 K1
x0
sDd
Dd
DTe 0
M
M
M
2:89
If rstly the contribution from the lower part of HeffronPhillips model, DTe , is
not considered, the electromechanical oscillation loop of generator shown in
Fig. 2.14 is described by the following second-order differential equation
s2 Dd
D
x0 K1
sDd
Dd 0
M
M
2:90
Equation (2.90) in fact is the linearized model of the single-machine innite-bus

power system when the dynamic of exciter and the AVR are not considered. This is
the case when the generator is modelled only by the rotor motion equation in
Eq. (2.3).
Ddt a0 e2Mt cos xNOF t b0
D
2:91
2
D 4x0 K1 .
where a0 and b0 are two constants and xNOF
M
M
Equation (2.91) describes the behaviour of rotor motion, i.e. the acceleration and
deceleration to store or release electric power. Hence, it determines the variations of
active power supplied by the generator during dynamic transient (electromechanical
transient), when the power system is subject to small disturbances. If 2D
M is small or
negative, a poorly damped or magnitude-increasing power oscillations occur. This
is the electromechanical oscillation associated with the rotor motion of synchronous
generator, i.e. the power system low-frequency oscillation.
In Eq. (2.91), xNOF is called the angular frequency of natural oscillation. The
angular oscillation frequency, xs , of power oscillation in the single-machine
innite-bus power system is normally very close to the angular frequency of natural
oscillation. Equation (2.91) indicates that the damping of power oscillation of the
single-machine innite-bus power system is determined by the coefcient of the
D.
rst-order derivative in the second-order differential equation in Eq. (2.91) M
At the angular oscillation frequency xs , the electric torque contributed from the
lower part of HeffronPhillips model can be decomposed into two components
1
2
DTe Td Dx Ts Dd
2:92
44
Equation (2.89) becomes

s2 Dd

D
Td
x0 K1 x0 Ts
sDd
Dd 0
M Mx0
M
M
2:93
Obviously, from the discussion on Eq. (2.91), it is easy to understand that the
component in the decomposition of DTe , Td Dx, contributes to the damping of
power oscillation. This component is called the damping torque. In Eq. (2.92), Ts Dd
is called the synchronizing torque.
2.3.1.2
Electric Torque Contributed from the PSS
From Fig. 2.10, it can be seen that the electric torque contributed from the lower
part of HeffronPhillips model is as follows:
DTet Fdelta sDd Fpss sDupss
2:94
where Fdelta s and Fpss s are the transfer function from Dd and Dupss respectively,
to form the electric torque contribution to the electromechanical oscillation loop of
generator.
The electric torque contribution from the PSS is as follows:
DTpss Fpss sDupss
2:95
Figure 2.15 shows that the PSS contributes the electric torque, DTpss , to the
electromechanical oscillation loop of generator. Obviously, Fpss s is the transfer
function of forward path from the stabilizing signal of the PSS to the
pss
K2
E q
KA
Td0's+K3
1+sTA
K6
Fig. 2.15 Forward path of stabilizing signal of the PSS
+ upss
45
electromechanical oscillation loop of generator in the HeffronPhillips model.

From Fig. 2.15, it can be obtained that
Fpss s K2
KA
1
K
K3 sT0d0 1 sTA
A
K2
2:96
0
K
1
A
1
sTA K6 KA
K
sT
3
1 K6
d0
0
K3 sTd0 1 sTA
Hence, at the angular oscillation frequency, xs , the electric torque provided by

the PSS to the electromechanical oscillation loop is as follows:
pss jxs Dupss
DTpss F
For example, if the PSS is a pure-gain controller and takes the deviation of rotor
speed of generator as the feedback signal, i.e. Dupss Kpss Dx, the electric torque
contributed from the PSS to the electromechanical oscillation loop of generator is as
follows:
pss jxs Dx Kpss ReF
pss jxs Dx jKpss ImF
pss jxs Dx 2:97
DTpss Kpss F

pss jxs and Im F

pss jxs denote the real and imaginary part of
where Re F
pss jxs , respectively (this notation will be used throughout the book). From the
F
rst equation in Eq. (2.41), it can have sDd x0 Dx, i.e.
Dx
jxs
Dd
x0
2:98
By substituting Eqs. (2.98) into Eq. (2.97), it can be obtained that

pss jxs Dx xs Kpss Im F

pss jxs Dd
DTpss Kpss ReF
x0
2:99
pss jxs Dx, and the

Hence, the damping torque supplied by the PSS is Kpss ReF
xs
pss jxs Dd.
synchronizing torque is x0 Kpss ImF
2.3.1.3
Damping Torque and Synchronizing Torque Derived

from the General Linearized Model
In Eq. (2.78) or Fig. 2.13, denote DTe aT238 Dx38 . From Eq. (2.78) or
Fig. 2.13, it can have
s2 Dd
a22
a21
sDd x0
Dd DTe 0
M
M
2:100
46
Taking the similar procedure of DTA presented above in Sects. 2.3.1.1 and
2.3.1.2, transfer function of the forward path of stabilizing signal of the PSS can be
obtained to be
Fpss s aT238 sI A33 1 bpss3
2:101
At the angular oscillation frequency, xs , the electric torque contribution from the
PSS is as follows:
DTpss aT238 jxs I A33 1 bpss3 Dupss
2:102
The electric torque can be decomposed into the damping and synchronizing
torque. The damping torque contribution from the PSS determines its effect on the
damping of power oscillation.
Assume that the installation of PSS brings about a change of damping coefcient
Dpss Dx in the electromechanical oscillation loop of generator. The state-space
representation of power system with the PSS installed can be equivalently written
as follows:
2
0
7 6 a21
6
4 sDx 5 4
M
Dx38
a138
sDd
x0
32
Dd
7
aT238 76
54 Dx 5
M
Dx38
A33
a22 Dpss
M
a238
2:103

i
@k
@Ac
Ti
vi
w
@Dpss
@Dpss

w
i1
i2
w
i2
w
vi2
M
@ 6
T
i3
w
4 a21
M
@Dpss
a138
x0
a22 Dpss
M
a238
32
3
vi1
7
aT238 76
54 vi2 5 2:104
M
vi3
A33
0
From Eqs. (2.83), (2.102), and (2.104), it can be seen that the residue in fact
measures the effect of the PSS on the electromechanical oscillation mode of the
i , it is equal to the forward path of the PSS
system. At the complex frequency k
multiplied by the sensitivity of the mode to the damping torque contribution.
47
2.3.2
Damping Torque Analysis and Design of PSS

by Phase Compensation
2.3.2.1
Theoretical Basis of the Damping Torque Analysis
This section explains the theoretical basis of the damping torque analysis by use of
the PhillipsHeffron model as follows.
Firstly, the effect of PSS is not considered, i.e. Dupss 0. From Fig. 2.10, it can
have
2

Ms Ds x0 K1 Dds x0 DTs
2:105
DTs Fdelta sDds
where Fdelta s is the transfer function from Dds to DTs. Combining two
equations above gives
2
Ms Ds x0 K1 x0 Fdelta s Dds 0
2:106
Thus, characteristic equation of the system is as follows:
Ms2 Ds x0 K1 x0 Fdelta s 0
2:107
Solutions of the characteristic equation are the eigenvalues of state matrix of

system model given by Eq. (2.42). One of the pair of complex solutions is called the
electromechanical oscillation mode. Its real part denes the damping of power
s n jxs . In the complex frequency domain, it
oscillation. Denote the mode as k
s
should have
2 Dk
s x0 K1 x0 F
s 0
delta k
Mk
s
2:108
The second equation in Eq. (2.105) expressed in the complex frequency domain
is as follows:
s Ddk
s
s F
delta k
DTk
2:109
Also in the complex frequency domain, the rst equation in Eq. (2.41) becomes
s
Dxk
ns jxs
s ns Ddk
s j xs Ddk
s
Ddk
x0
x0
x0
2:110
Let the electric torque dened by Eq. (2.109) be decomposed as follows:

s Ts1 Ddk
s Td1 Dxk
s
DTk
2:111
48

s Ddk
s Ts1 Ddk
s Td1 ns Ddk
s jT xs Ddk
s
delta k
F
d1
x0
x0
2:112
That is
s Ts1 Td1 ns jT xs
delta k
F
d1
x0
x0
2:113

(
s
delta k
Td1 xx0s ImF
s Td1 ns
delta k
Ts1 ReF
2:114
x0
The above derivation indicates that in the complex frequency domain, the
electric torque can be decomposed into damping and synchronizing torque
according to Eq. (2.111). Substituting Eqs. (2.111) into (2.108), it can have
2 Dk
s x0 K1 Ddk
s x0 Ts1 Ddk
s x0 Td1 Dxk
s
Mk
s
s Ddk
s Td1 k
s
x0 Ts1 Ddk
2:115
Thus,
s x0 K1 x0 Ts1 0
2 D Td1 k
Mk
s
2:116
Solution of the above equation is as follows:

ns
D Td1
2M
2:117
Equation (2.117) indicates that the damping torque affects the real part of
electromechanical oscillation mode, i.e. the damping of power oscillation.
2.3.2.2
Graphical Explanation of the Damping Torque Analysis
Equations (2.16) and (2.39) include the following linearized rotor motion equation
of generator
:
D d x0 Dx
1
:
D x DPt DDx
M
2:118
49
Without referring to any linearized model of single-machine innite-bus power

system, assume DPt being comprised of contributions from Dd and stabilizing
signal of the PSS, Dupss . That is to let
DPt DPt Dd DPt Dupss
DPt Dd Fdelta sDd; DPt Dupss Fpss sDupss
2:119
At a known angular oscillation frequency xs , if DPt is decomposed in the

complex frequency domain, then
delta jxs Dd Cdelta Dd Ddelta Dx;
DPt Dd F
pss jxs Dupss Cpss Dd Dpss Dx
DPt Dupss F
2:120
Obviously, Ddelta Dx and Dpss Dx are the damping torque contributed to the
electromechanical oscillation loop of generator from Dd and Dupss , respectively.
Function of the damping torque component in suppressing the power oscillation can
be explained graphically by the use of the linearized equal-area criterion as follows.
Without affecting the conclusions of following discussion,
it is assumed

that in
Eq. (2.120), Ddelta 0. When there is no PSS installed DPt Dupss 0 ,
DPt DPt Dd Cdelta Dd
2:121
The linearized DPt Dd curve is line af as shown in Fig. 2.16. In Fig. 2.16, the
operating point of system at steady state is d (Pt0 ; d0 ) and it moves to point a
(Pt1 ; d1 ) after the system is subject to a small disturbance. Hence, when the operating point moves down from the initial point a (Pt1 ; d1 ) along line af, it will stop at
point f (Pt20 ; d20 ) with area adc being equal to area dgf. Obviously, in this case,
jPt1 Pt0 j jPt20 Pt0 j; jd1 d0 j jd20 d0 j, power oscillation is of unchanged
magnitude and not damped at all.
When the PSS is installed to provide a pure positive damping torque,
DPt Dupss Dpss Dx assuming Dpss [ 0
2:122
From Eqs. (2.120), (2.121), and (2.122), it should have

DPt DPt Dd DPt Dupss Cdelta Dd Dpss Dx;
2:123
When the operating point moves down from point a (Pt1 ; d1 ), power angle of
generator decreases and thus Dx\0. Dpss Dx\0 is added on Cdelta Dd as shown in
Eq. (2.123). Hence, the operating point should move below line af along curve
DPt Cdelta Dd Dpss Dx;. When the operating point stops moving, Dx 0. Thus,
it should stop on line af at point c (Pt2 ; d2 ) with area A1 being equal to area A2 .
50
Fig. 2.16 Graphical
explanation of the DTA
Pt
Pt=Cdelta+Dpss
Pt=Cdelta
A1
Pt0
d
A2
c
Direction of
< 0
2'
Time response
with the PSS
time
Time response without the PSS
Obviously, jPt2 Pt0 j\jPt1 Pt0 j; jd2 d0 j\jd1 d0 j, indicating extra positive
damping is provided by the PSS to the power oscillation. A similar analysis can be
carried out to examine the case when the operating point moves up from point c
(Pt2 ; d2 ).
The above discussion explains the function of damping provided by the PSS in
suppressing the power oscillation. It is important to note that the explanation relies
only on the linearized rotor motion equation in Eqs. (2.118), (2.119), and (2.120)
without referring to any particular type of model of power system. This means that
for any type of linearized model of power system, including that of a multi-machine
power system, if Eqs. (2.119) and (2.120) can be established on the basis of the
model, the above procedure can be applied.
2.3.2.3
Design of PSS by the Phase Compensation Method [4]
If the rotor speed of generator is taken as the feedback signal of the PSS, transfer
function of the PSS is Tpss s, that is
51
DuPSS Tpss sDx
2:124
From Eqs. (2.95) or (2.102) and above equation, it can have

DTpss Fpss sTpss sDx
2:125
At the angular oscillation frequency, xs , the decomposition of the electric torque

contributed by the PSS is as follows:
pss jxs T
pss jxs Dx
DTpss F

pss jxs T
pss jxs Dx jImF
pss jxs T
pss jxs Dx
Re F
pss jxs T
pss jxs Dx xs ImF
pss jxs T
pss jxs Dd
ReF
x0
Tpssd Dx Tpsss Dd
2:126
The damping and synchronizing torque provided by the PSS is Tpssd Dx and
Tpsss Dd, respectively. In order to achieve the most efcient design, ideally the PSS
should provide only the damping torque, that is,
DTpss Dpss Dx;
Dpss [ 0
2:127
where Dpss is the coefcient of the damping torque which needs to be provided by
the PSS. Hence, from Eqs. (2.126) and (2.127), it can be seen that design of the PSS
should satisfy that
pss jxs T
pss jxs
Dpss F
2:128
According to Eq. (2.128), design of the PSS should set the phase of the PSS,
pss jxs , to be equal to the minus phase of the forward path, \F
pss jxs , that is to
\T
design the PSS such that it can compensate the phase lag of the forward path and
ensure it to provide a pure positive damping torque. Hence, the method to design
the PSS based on Eq. (2.128) is called the phase compensation method.
If it is denoted that

pss jxs Fpss \/; T
pss jxs Tpss \c
F
2:129
The phase compensation method requires

Tpssd Fpss Tpss cos/ c Dpss

Tpsss Fpss Tpss sin/ c 0
2:130
52
This can be achieved by setting

Dpss
c /; Tpss
Fpss
2:131
Often the PSS is constructed as a leadlag block with its main part of transfer
function to be
Tpss s Kpss
1 sT2 1 sT4
1 sT2
1 sT4
Kpss1
Kpss2
1 sT1 1 sT3
1 sT1
1 sT3
2:132
where Kpss Kpss1 Kpss2 . Parameters of the PSS then can be set to satisfy
Dpss
1 jxs T2
/
\
1 jxs T1 Fpss
2
1 jxs T4
/
1:0\
Kpss2
1 jxs T3
2
Kpss1
2:133
for the PSS to provide a positive damping torque Dpss Dx.

With the PSS installed, from Fig. 2.10, it can have
Ms2 Ds x0 K1 Dds x0 DTdelta s x0 DTpss s
DTdelta s Fdelta sDds
DTpss s Fpss sTpss s)Dxs
2:134
c n jxc be the electromechanical oscillation mode of the closed-loop

Let k
c
system with the PSS installed. According to the DTA discussed above in
Sect. 2.3.2.1, in the complex frequency domain, it can have
c sDdk
c F
c C1 Ddk
c D1 Dxk
c
delta k
DTdelta k
c F
c T
c Dxk
c Cpss Ddk
c Dpss Dxk
c
pss k
pss k
DTpss k
2:135
Taking the similar procedure of discussion in Sect. 2.3.2.1, it can easily prove
c affects the damping of the electromechanical oscillation mode.
that Dpss Dxk
Hence, the PSS can be designed to satisfy
c Tpss k
c Dpss
Fpss k
2:136
c n jxc is given, the PSS can be

When the PSS is being designed, if k
c
designed by using the phase compensation method dened by Eq. (2.136) to move
c n jxc in
the electromechanical oscillation mode strictly to the position at k
c
the complex plane. If only the amount of damping torque provision
Dpss Dx; Dpss [ 0 is given, the PSS can also be designed by the phase compensation
method introduced above from Eqs. (2.126) to (2.133).
2.4 Examples
2.4
53
Examples
2.4.1
Linearized Mathematical Models of an Example

Power System
2.4.1.1
Linearized Mathematical Model with Full Model

of Generator Used
Parameters (in p.u.) of an example single-machine innite-bus power system are as

follows:
1. Generator:
xd 1:18; xq 0:78; xad 1:0; xaq 0:6; xD 1:11; xQ 0:73; xf 1:13;
ra 0:005; rf 0:00075; rD 0:002; rQ 0:04; M 7 s; D 0; T0d0 5 s
2. AVR: KA 100; TA 0:01 s
3. Transmission line: xt 0:15
4. Steady-state operating point: Pt0 0:5; Vt0 1:05; Vb0 1:0
At the steady-state operating point, the complex power received at the innite
busbar is as follows:

t0 V
b0
V

Vb0
Pt0 jQb0
jxt
where Qb0 is the reactive power received at the innite busbar and * denotes the
b0 as the reference phasor, that is,
conjugate of a complex phasor. By choosing V
b0 V
Vb0 , from the above equation, it can be obtained that
V
b0
Qb0
2s
3

2
Vb0 4
xt Pt0
V2t0
Vb0 5 0:3155 p:u:
xt
Vb0
In the above, calculation is in p.u. Throughout the following calculation in p.u.,

p.u. is omitted.
The line current is as follows:
It0 Pt0 jQb0 0:5 j0:3155
Vb0
54
Terminal voltage of generator is :

t0 jxtIt0 Vb0 1:0473 j0:075
V
At steady state, iD0 iQ0 0; x0 1, Eqs. (2.1) and (2.2) become
0 vtd0 ra id0 wq0
0 vtq0 ra iq0 wd0
0 vf0 rf if0

xd xad id0
wd0
wf0
xad xf
if0
wq0 xq iq0
Thus, according to Eq. (2.28)

vtd0 ra id0 wq0 ra id0 xq iq0
vtq0 ra iq0 wd0 ra iq0 xad if0 xd id0 ra iq0 Eq0 xd id0
Hence, in dq coordinate of generator,
t0 Vtd0 jVtq0 ra id0 xq iq0 jra iq0 Eq0 xd id0
V
ra id0 jiq0 xq iq0 jxq id0 jEq0 xd id0 xq id0
Let an imaginary electromotive force (EMF) be
EQ0 Eq0 xd id0 xq id0 Eq0 xd xq id0
Thus,
t0 ra jxq id0 jiq0 E
Q0
V
Q
Thus, the q-axis of generator can be found by calculating the imaginary EMF E
Q0 V
t0 ra jxq It0 1:2959 j0:4634 1:38\19:68
E
Hence, d0 19:68 . From the above equation about Eq0 and EQ0 , it can have
Eq0 EQ0 xq xd id0 1:1506
Eq0
if0
1:5624
xad
2.4 Examples
55
Fig. 2.17 Phasor diagram for

computing initial values of
state variables
EQ
[ra + j(x q x d )]I t
Eq
jx d I t
v tq
jx t I t
Vt
Vb
It
vtd
Above computation can be shown by the phasor diagram of Fig. 2.17 where
subscript 0 is omitted. From Fig. 2.17, it can be seen that
id0 It0 sind0 u0 0:4654; iq0 It0 cosd0 u0 0:3646
vd0 Vb0 sin d0 0:3368; vq0 Vb0 cos d0 0:9416
Because
t0 vtd0 jvtq0 jxtIt0 Vb0 jxt id0 jiq0 vd0 jvq0
V
thus
vtd0 vd0 xt iq0 0:2820
vtq0 xt id0 vq0 1:0114
if0
Eq0
1:5624
xad
56
Because iD0 0; iQ0 0, thus according to Eq. (2.2), it can be obtained that
2
wd0
xd
xad
xad
32
id0
1:0132
7
7 6
76
7 6
6
4 wf0 5 4 xad xf xad 54 if0 5 4 1:3001 5
1:0970
wD0
xad xad xD
iD0
"
#

wq0
xq xaq iq0
0:2844
wQ0
xaq xQ
iQ0
0:2187
Denote
2
xd
6
4 xad
xad
xad
xf
31 2
xad
a11
6
7
xad 5 4 a21
a12
a22
xad
xD
a31
a32
xq
xaq
b11
b12
xaq
xQ
b13
b14
1
3 2
4:2331
a13
7 6
a23 5 4 1:8311
1:8311
5:1570
2:1640 2:9963
a33

3:6842 2:8653
2:8653 3:7249
3
2:1640
7
2:9963 5
5:5498

2
Agdq
x0
D
M
0
6
60
6
60
6
6
60 x w
0 q0
6
6
6 0 x0 wd0
6
60
0
6
6
40
0
2
0
0
0
314:16
60
0
6
6
60
0
6
6 0 89:3377
6
6
6 0 318:3167
6
60
0
6
6
40
0
0
Vtd0
a11M
b V
11M tq0
0
Vtd0
a12M
Vtd0
a13M
b V
12M tq0
0
T1
a11 x0 ra
x0 x
a12 x0 ra
a13 x0 ra
x0 x
b11 x0 ra
b12 x0 ra
x0
0
a21 x0 rf
a31 x0 rD
0
0
a22 x0 rf
a32 x0 rD
a23 x0 rf
a33 x0 rD
0
0
b21 x0 rQ
0
0
0
0:1706
0:5037
0:0738
100
0
0
6:6494
0
314:16
0
2:8762
0
314:16
314:16
0:4314
5:4760
0
0
1:2151
1:35797
1:8826
36:0068
3
7
7
7
7
7
7
7
7
7
7
7
7
7
7
5
b22 x0 rQ
3
0
0:0872 0:4140 7
7
7
7
0
0
7
7
3:3992
0
7
7
0
4:5008 7
7
7
0:7060
0
7
7
5
3:4871
0
0
0
46:8088
2.4 Examples
57
3 2
3
0
0
0
0
iq0 7 6
6 id0
M 7
6
0:0665 0:0521 7
M
7
7 6
6
6
6 KA Vtd0
KA Vtq0 7 6 2686:1 9632:5 7
7
6 T V
T V 7 6
7
A
t0
A
t0 7
6
7
6
7 6 314:1593
6 x
0
7
0
7
6
0
6
7
6
7 6
7
0
314:1593
7
6
0
x
7
6
0
7 6
6
7
7 6
6
0
0
7
0
0
7 6
6
7
7 4
6
5
0
0
5
4
0
0
0
0
0
0

0 0 0 a11 0 a12 a13 0
0 0 0 0 b11 0
0 b12

0 0 0 4:2331
0
1:8331 2:1640
0
0 0 0
0
3:4862
0
0
2:8653

0 xt
0
0:15
0:15
0
xt
0

Vb cos d0 0
0:9416 0
Vb sin d0 0
0:3367 0
2
Bgdq
Cgdq
Fdq1
Fdq2
Thus, state matrix and control vector of linearized state-space model are obtained
to be
Agcdq Agdq Bgdq Fdq1 Cgdq Bgdq Fdq2
2
0
314:16
0
0
6 0:045
0
0
0:20
6
6
6 714:3
0
100 6116:3
6
6 295:8
89:34
0
6:64
6
6
6 105:78 318:3
0
513:64
6
6
0
0
314:2
0:43
6
6
4
0
0
0
1:36
0
3
bpss
0
3
2
2
0
0
6 0 7 6 0 7
7
6 7 6
7
6 KA 7 6
6 T 7 6 10000 7
7
6 A7 6
6 0 7 6 0 7
7
6 7 6
6 76
7
6 0 7 6 0 7
7
6 7 6
6 0 7 6 0 7
7
6 7 6
7
6 7 6
4 0 5 4 0 5
0
0
0
0:47
0:09
1404:6 2646
0
0:104
3127
478:4
5:48
0
2:88
86:29
1:22
3:39
101:9
0:70
0
36:0
1:88
0
3:48
0
0
0:38
1154:5
7
7
7
7
7
135:03 7
7
7
4:50 7
7
7
0
7
7
5
0
46:8
58
Eigenvalues of above state matrix are computed to be

1 ;k
2 24:14 j971:56
k
3 ;k
4 36:33 j359:63
k
k5 35:57
k6 ; k
7 0:71 j8:44
k8 5:71
According to Eq. (2.63), the oscillation frequency f i (Hz) and damping fi for the
6 and k7 , are as follows:
electromechanical oscillation mode, k
xi
1:34 Hz
2p
ni
fi q
0:0838;
2
ni x2i
fi
2.4.1.2
i 6; 7
HeffronPhillips Model of Example Power System
With D and Q damping winding of generator being ignored, from the given
parameters of above example power system and results of calculation, it can have
E0q0
x2
x0d xd ad 0:2951
x

f
0
EQ0 xq xd itd0 1:1506

K1 1:5248; K2 0:7602; K3 2:9885;
K4 0:6727; K5 0:0027; K6 0:3245
State equation of (2.41) can be
: 3
2
Dd
0
314:16
:
7
6
0:2178
0
6 D x0 7 6
6 _ 76
0
4 DEq 5 4 0:1346
0
26:4729
0
DE_
2
fd
obtained to be
32
3
3 2
Dd
0
0
0
7
7 6
6
0:1086
0 7
76 Dx0 7 6 0 7Dupss
5
5
4
4
DEq
0:5978 0:2
0 5
0
DEfd
3245 100
10000
Eigenvalues of state matrix are as follows:

2;3 0:0114 8:2610; k4 7:6008
k1 92:9741; k
2.4 Examples
59
According to Eq. (2.63), oscillation frequency f i (Hz) and damping fi for the
2;3 , are as follows:
electromechanical oscillation mode, k
xi
1:315 Hz
2p
ni
fi q
0:0014;
2
ni x2i
fi
2.4.2
i 2; 3
Modal Analysis of Example Power System
HeffronPhillips model established above is used to demonstrate the modal analysis

of example power system in this section.
2.4.2.1
Modal Decomposition and Stability of Example Power System
Right eigenvectors corresponding to each of eigenvalues of state matrix of Heffron

Phillips model are calculated to be
2
8:4781 106
6 2:5091 106
6
v1 6
4
0:0022
0:9014
7
6
7
j0:0237
7
6
7
7
7; v2 6
4 0:0050 j0:0050 5
5
0:3868 j0:1930
3
2
0:0077
7
6
6 0:0002 7
7
7
6
6

v3 6
7; v4 6
7
4 0:0050 j0:0050 5
4 0:0284 5
1
0:9014
j0:0237
0:3868 j0:1930
0:9996
Hence,
v1
v2
v3 v4
V
2
8:4781 106
6 2:5091 106
6
6
4
0:0022
1
0:9014
j0:0237
0:0050 j0:0050
0:9014
j0:0237
0:0050 j0:0050
0:0077
0:0002
0:0284
0:3868 j0:1930
0:3868 j0:1930
0:9996
3
7
7
7
5
60
From Eq. (2.53),

T w1
W
2
2
w

w4 T V
3
w
1
0:5554 j0:0001
0:5554 j0:0001
0:2524
0:1773
6 0:8529 0:0312 j21:1226 0:0312 j21:1226 7:3269 7

7
6
6
7
4 38:0214
38:2206 5
0:1626 j0:1475
0:1626 j0:1475
0:0827
0:0003 j0:0003
0:0003 j0:0003
1:0823
Modal decomposition thus is obtained to be
szi ki zi ; wTi bo u,
i 1; 2; 3; 4

Without considering the PSS Dupss 0 , solution of modal decomposition is
obtained to be
zi t) zi 0eki t ;
i 1; 2; 3; 4

Z0 V1 X0 WT X0
Time response of state variables can be written as follows:
xk t)
4
X
vki zi 0eki t
i1
4
X
vki zi 0eni t cos xi t j sin xi t
i1
It can be seen that when time approaches innity t ! 1,

4
X
lim xk t) lim
t!1
t!1
i1
92:9741t
lim fe
t!1
vki zi 0eki t lim
4
X
vki zi 0eni t cos xi t j sin xi t

i1
vk1 z1 0 e0:0114t vk2 z2 0cos8:2610t j sin8:2610t
t!1
e0:0114t vk3 z3 0cos8:2610t j sin8:2610t e7:6008t vk4 z4 0g 0

Since Xt Ddt Dxt
DE0q t DE0fd t
and
Ddt dt d0 ; Dxt xt x0 ;
DE0q t E0q t E0q0 t; DE0fd t E0fd t E0fd0 t
2.4 Examples
61
lim xk t) 0;
k 1; 2; 3; 4 should give
t!1
lim dt d0 ; lim xt x0 ; lim E0q t E0q0 t; lim E0fd t E0fd0 t
t!1
t!1
t!1
t!1
All the state variables return to their initial points X0, the equilibrium point of
the system. Hence, the system is stable in terms of small-signal stability.
2.4.2.2
Modal Analysis of the AVR
State matrix, control vector, and output vector of Eq. (2.71) are as follows:
"
#

0
w0
0
314:16
A0
D 0:2178
1
K
0
M M
"
#

0
0
b0
cT0 1 0
K
2
M
0:1086
1;2 j8:2725 with the corresponding
Its eigenvalues are calculated to be k
eigenvectors to be

v1

0:9997
0:9997
; v2
j0:0263
j0:0263
Thus
T w
1 v1 v2 1
1 w
2 T V
W

0:9997
0:9997 1
0:5002
j0:0263 j0:0263
j18:9947
0:5002
T
j18:9947
Hence, the controllability and observability index, respectively, can be obtained

to be
T1 b0
w
T2 b0
w
cT0
v1
cT0
v2

0
j2:0628
0:5002 j18:9914
0:1086

0
j2:0628
0:5002 j18:9914
0:1086

0:9997
1 0
0:9997
j0:0263

0:9997
0:9997
1 0
j0:0263
62
The residue is as follows:

1 w
T1 b0 cT0 v1 j2:0622
R
2 w
T2 b0 cT0 v2 j2:0622
R
From Eq. (2.77), when a KA , it can have
TA ki 1K4 KA K5
@ 0
i
i 1 KA K6
@k
ki Tdo K3 TA k
Ri
@KA
@KA
i 1 KA K6 TA k
i 1K4 KA K5 K6
i T0 K3 TA k
K5 k
do
Ri
i 1 KA K6 2
i T0 K3 TA k
k
do
0:00018158 j0:00005839
i 1; 2
When a TA it can have

@ 0TA ki 1K4 KA K5
i
@k
ki Tdo K3 TA ki 1 KA K6
Ri
@TA
@TA
i K4 k
i 1 KA K6 TA k
i 1K4 KA K5 k
i T0 K3 TA k
i T0 K3 k
i
k
do
do
Ri
2
0
i 1 KA K6
i T K3 TA k
k
do
0:0357 j0:1533
i 1; 2
Above results indicate that increase of the gain value and time constant will
move the electromechanical oscillation mode towards the right on the complex
plane and hence is detrimental to the small-signal angular stability of the power
system. It has been well known that the fast-acting high-gain AVR may be detrimental to the damping of power system electromechanical oscillation modes. This
means that increase of gain value of the AVR could move the oscillation mode to
the right. However, increase of the time constant (slower action of the AVR) should
not.
In order to further clarify the results of derivative of the oscillation mode in
respect to the parameters of the AVR obtained above, Fig. 2.18 presents the loci of
movement of the electromechanical oscillation mode on the complex plane with the
change of gain value and time constant of the AVR. In Fig. 2.18, KA increases from
KA 50 to KA 100 with TA 0:01 and TA increases from TA 0:01 to TA
0:1 with KA 100. From Fig. 2.18, it can be seen that although at the point
KA 100; TA 0:01 where the derivatives are calculated, the trend of loci
movement is towards the right with the increase of the gain value and time constant,
and the oscillation mode in fact moves towards left when the time constant of the
AVR increases. The trend of the loci with the increase of the time constant of the
AVR actually changes the direction at the point KA 100; TA 0:01.
2.4 Examples
63
Fig. 2.18 Loci of the movement of the oscillation mode on the complex plane with the changes of
parameters of the AVR
2.4.2.3
Design of the PSS by Pole Assignment for the Example Power

System
A PSS can be installed to increase the damping of example power system by

assigning the electromechanical oscillation mode, k2;3 0:0114 j8:2611, to a
target position on the complex plane, kc 0:9 j8:2611, which is damping over
10 %.
Let the transfer function of the PSS be (see Eq. (2.88)) with T 0:1 s
Tpss s Kpss
1 saT2
1 sT2
64
State-space model of example power system with the PSS to be installed is

2
Dd_
6
6 Dx_
6
6 DE_ 0
q
4
0
DE_
2
0
314:16
7 6
7 6 0:2178
0
76
7 4 0:1346
0
5
fd
Dx 0
26:4729
0
0:1086
0
2
Dd
6 Dx
6
0 6
4 DE0q
0
0
32
Dd
0
0
76 Dx 7 6
7
76
7
7 6
76
6
7Dupss
7
0:5978 0:2 54 DE0q 5 4 0 5
DE0fd
3245 100
10000
3
7
7
7
5
DE0fd
Dupss Tpss sDx
Thus, the transfer function of open-loop system can be obtained to be
Gs cT sI A1 b
1:1s4
110:6s3
238:8s
854:5s2 7563s 53020
From the characteristic equation of closed-loop control system of Eq. (2.85), for
c , it should have
the electromechanical oscillation mode k
c
Tpss k
1
1
c 0:0753 j0:0996
Gk
That is
Kpss
c aT2
1 k
1
2

0:0753 j0:0996
1 kc T
By solving the above equations, parameters of the PSS can be obtained to be

Kpss 2:6151
a 2:4256
In order to establish the state-space model of closed-loop system with the PSS
installed, let
1 saT
Dx
1 sT
1 saT
Dx1
Kpss
1 sT
Dx1
Dupss
2.4 Examples
65
Thus, state equation of the PSS is as follows:

1
1
Dx a sDx Dx1
T
T

K1
1 D
K2
1
a Dx
aDE0q Dx1
aDd
T M
T
M
M
1
1
Kpss Dx1 Dupss Kpss a sDx1
T
T
K1 2
1 D
K2 0
a Dd Kpss a aDx Kpss a2
DEq
Kpss
T M
M
M
1
1
1 a Kpss Dx1 Dupss
T
T
sDx1
sDupss
By writing the state equation of open-loop system and the PSS together, the
state-space model of closed-loop system is obtained to be
2
2
0
7 6
6
7 6 0:218
6
7 6
6
6
q 7
6 0:134
6
0 7 6 26:582
6 DE_ 7 6
fd 7
6
4 Dx_ 1 5 4 0:528
3:352
Du_ pss
:
Dd
:
Dx
0
DE_
314:159
0
0
0:109
0
0:597
0
3245:044
10
0:263
63:432
1:671
0
0
0:200
100
0
0
32
3
Dd
0
0
7
6
0
0 7
76 Dx0 7
6 DEq 7
0
0 7
76
7
0 7
6
0
10000 7
76 DEfd 7
10
0 54 Dx1 5
Dupss
37:281 10
Eigenvalues of state matrix are calculated to be

1 93:4535
k
2;3 0:8995 j5:2621
k
4;5 9:4136 j5:0918
k
6 6:5176
k
Hence, the electromechanical oscillation mode is successfully assigned to the
target position.
Figure 2.19 shows the simulation result of example power system without and
with the PSS installed. At 1.0 s of simulation, a three-phase to-earth short circuit
occurred on the transmission line which was cleared in 100 ms. From Fig. 2.19, it
can be seen that the low-frequency oscillation is damped effectively by the PSS
designed by use of the method of pole assignment.
66
rotor angle (degree)
50
with PSS installed
without PSS
40
30
20
10
-10
10
time (second)
Fig. 2.19 Simulation result of example power system without and with PSS installed
2.4.3
Damping Torque Analysis of Example Power

System
HeffronPhillips model established is again used in this section to demonstrate the

DTA of example power system.
2.4.3.1
Damping Torque Provided by the AVR in the Example Power

System
From Fig. 2.10, it can be seen that the electric torque provided by the AVR to the
electromechanical oscillation loop of generator is as follows:
DTavr
KA
1
sT0do K3 sTA 1
K2
Dd
1 0 K6 sTKA 1
sTdo K3 A
K2 K5 KA
Dd Favr sDd

K6 KA sT0do K3 sTA 1
K5
At the complex frequency of the electromechanical oscillation with

s n jxs k1 0:0114 j8:2610, let the decomposition of the electric
k
s
torque provided by the AVR be
2.4 Examples
67

s Tsavr Dd k
s Tdavr Dx k
s
DTavr k
Because
s
Dxk
ns jxs
s ns Ddk
s j xs Ddk
s
Ddk
x0
x0
x0
thus
xs
s Tsavr Ddk
s Tdavr ns Ddk
s jT
s
DTavr k
Ddk
davr
x0
x0
That is
Tdavr
xs
s
ImFavr k
x0
Since
K2 K5 KA

s TA 1
K6 KA ks T0do K3 k
K2 K5 KA

2
s T0 TA k
K6 KA K3 K3 TA T0d0 k
d0
s
K2 K5 KA

K6 KA K3 K3 TA T0d0 ns jxs T0d0 TA ns jxs 2
K2 K5 KA
K2 K5 KA
2

a jb 0:0024 j0:0031
a jb
a b2
s
Favr k
where

a K6 KA K3 K3 TA T0d0 ns T0d0 TA n2s x2s 31:9695

b K3 TA T0d0 xs 2T0d0 TA ns xs 41:5426
thus it can be obtained
x0
s x0 K2 K5 KA b
ImFavr k
xs
xs a2 b2
0:0988
Tdavr
68
Since
@a
@ KA
K6 0:3245;
@a
@ TA

K3 ns T0d0 n2s x2s 341:2561
@b
@ KA
0;
@b
@ TA
K3 xs 2T0d0 ns xs 23:7488
sensitivity of the damping torque provided by the AVR to its parameters can be
obtained to be
h
@Tdavr
@KA
@Tdavr
@TA
x0 K2 K5
xs

i

b KA @@Kb a2 b2 KA b 2a @@Ka 2b @@Kb
A
A
0:0003
a2 b2 2
h

i
2
@b
@a
@b
2
x0 K2 K5 KA @ TA a b b 2a @ TA 2b @ TA
0:9055
xs
a2 b2 2
A
The above results indicate that (1) with the increase of the AVR gain, less
damping torque will be provided by the AVR, detrimental to the system
small-signal angular stability and (2) with the increase of the AVR time constant,
less damping torque will be provided by the AVR, also detrimental to the damping
of low-frequency power oscillations.
2.4.3.2
Design of PSS Installed in the Example Power System

by the Phase Compensation Method
The PSS to be designed is to provide a damping torque DTpss Dpss Dx; Dpss 15.
From Eq. (2.96), the forward path of stabilizing signal of the PSS can be obtained to
be jxs j8:44
Fpss jxs K2
KA
0
K3 jxs Td0 1 jxs TA K6 KA
0:8598 j1:1451 1:4320\53:0989

The PSS adopts the deviation of rotor speed of generator as the feedback signal,
and its transfer function is as follows:
Tpss s K1
1 sT2 1 sT4
K2
with T1 0:09 s; T3 0:09 s
1 sT1 1 sT3
2.4 Examples
69
According to Eq. (2.133), parameters of the PSS are set to compensate the phase
of the forward path and thus obtained to be
Kpss Kpss1 Kpss2 3:2271; T2 0:2405 s; T4 0:2405 s
Let
1 sT4
Dx 9:3706 j4:6819Dx
1 sT3
1 sT2
Dx1 0:8946 j0:4470Dx1
Kpss1
1 sT1
Dx1 Kpss2
Dupss
Thus, state-space realization of the PSS is as follows:

Kpss2
1
Dx1
T4 sDx Dx
T3
T3

Kpss2 T4
1
Dx1
K1 Dd K2 DE0q DDx Dx
T3
T3 M

Kpss2 T4 K1
Kpss2
Kpss2 T4 K2 0
T4
1

Dd
DEq Dx1
1 D Dx
T3
T3 M
T3
M
T3 M
sDx1
0:3230Dd 6:1672Dx 0:1224DE0q 11:1111Dx1

Kpss1 T2 Kpss1
Kpss1 T2
Kpss2 T4 K1
1
Dd
Dupss

Dx1
T1
T1
T3 T1
T1
T3 M

Kpss2 T4 K2 0 Kpss2
T4

DEq
1 D Dx
T3 M
T3
M

Kpss T2 T4 K1
Kpss T2
Kpss T2 T4 K2 0
T4

Dd
DEq
1 D Dx
T1 T3 M
T1 T3
M
T1 T3 M

Kpss1 T2 Kpss1
1
+

Dx1 Dupss
T1
T1
T3 T1
sDupss
5:0182Dd 95:8026Dx 2:5018DE0q 108:0008Dx1 11:1111Dupss

By writing the state equation of open-loop system and the PSS together, state
matrix of closed-loop system is obtained to be
70
0
x0
K1
6
D

6
M
M
6
6
0
K0 4
6
6
Td0
6
A6
0
KTA K5
6
A
6

6
Kpss2
4
6 Kpss2 T4 K1
1 TTM
6
T
T
3M
3
3
4

T2 1 T4 D K
2 T 4 K1
K
T
PSS
PSS
T1 T3 M
T1 T3
M
2
0
314:16
0
0
6 0:2178
0
0:1086
0
6
6
6 0:1345
0
0:5977
0:2
6
6 26:5821
0
3245
100
6
6
4 0:3230 6:1672 0:1224
0
5:0182
95:8026
2:5018
0
2
K
M
K
03
Td0
KTA K6
A
20 T4 K2
KT
K2
3M
2 T 4 K2
T
T1 T3 M KPSS
0
0
0
0
11:1111
108:0008
T0d0
A
K
TA
T1
3

Kpss1
T2
1

T1
T3
3
0
0
0
0 3
0 7
7
7
7
0 7
7
7
KA 7
TA 7
7
7
0 7
7
5
1
T
1
7
7
7
7
0
7
10000 7
7
7
5
0
11:1111
Eigenvalues of state matrix are calculated to be

k1 93:6838
2;3 1:2125 j8:0051
k
k4;5 8:7171 j6:2080
k6 9:2771
Hence, damping of the electromechanical oscillation mode is successfully
increased by the PSS designed above via the phase compensation method.
Figure 2.20 shows the simulation result of the example power system without
and with the PSS installed. At 1.0 s of the simulation, a three-phase to-earth short
Fig. 2.20 Simulation result of power system without and with the PSS designed by use of the
phase compensation method
2.4 Examples
71
circuit occurred on the transmission line which was cleared in 100 ms. From
Fig. 2.20, it can be seen that the low-frequency oscillation is damped effectively by
the PSS designed by use of the phase compensation method.
2.4.3.3
Theoretical Basis and Graphical Explanation of the Damping

Torque Analysis
From Fig. 2.10 and Eq. (2.105), it can be obtained that

0
B
Fdelta s K2 @
K4
0
do
1
K3
sT
K
1 0 6 sTKA 1
sTdo K3 A
1
KA
sTA 1C
sT
K
KA A
6
1 0
sTdo K3 sTA 1
K5
0
do
1
K3
K2 K4 sTA 1 K5 KA

K6 KA sT0do K3 sTA 1
At the complex frequency of electromechanical

s n jxs k1 0:0114 j8:2610,
k
s
oscillation
with

delta ks 0:0042 j0:0042
F
According to Eq. (2.114)
(
s 0:1591
delta k
Td1 xx0s Im F

s Td1 ns 0:0042
delta k
Ts1 Re F
x0
Substituting the above result into Eq. (2.116), it can have

2 0:1553k
s 477:7128 0
7k
s
Solution of above equation is ks 0:0114 j8:2610. It is the electromechanical oscillation mode of example power system without the PSS being
installed.
s Dd K1 Dd Ts1 Dd Td1 Dx 1:5207Dd 0:1553Dx
DPt K1 Dd Fdelta k
This is the rst equation in Eq. (2.120). Figure 2.21 presents the Pt d curve
from simulation. At 1 s of simulation, the mechanical power input to the generator
72
without PSS
with PSS installed
Fig. 2.21 Pt d curve, corresponding variation of rotor angle and power output of the generator
of example power system
increased by 1 % and then returned to its original value in 10 ms. Pt d curve

depicted in Fig. 2.20 is the rst cycle of power oscillation starting from its rst
peak. Dashed curve is the case without the PSS installed.
With the PSS designed in Sect. 2.4.3.2 being installed,
c Dd F
c T
c Dxk
c
pss k
pss k
DPt K1 Dd Fdelta k
c
K1 Dd Ts1 Dd Td1 Dx Dpss Dxk
1:5207Dd 15:1553Dx
where kc 1:2279 j8:0264 is the electromechanical oscillation mode of
example power system with the PSS being installed. Solid Pt d curve in Fig. 2.21
is the case that the PSS is installed in the example power system. Figure 2.21
conrms the graphical explanation of the DTA illustrated in Fig. 2.16.
2.4 Examples
2.4.4
73
Equivalence Between the Damping Torque and Modal

Analysis
i , the residue is equal to

It is concluded in Sect. 2.3.1.3 that at a complex frequency k
the forward path of the PSS multiplied by the sensitivity of the mode to the
damping torque contribution. In this section, this conclusion is to be demonstrated
by example power system.
2.4.4.1
Demonstration by Use of HeffronPhillips Model of Example

Power System
In Sect. 2.4.2.3, state-space model of the example power system is obtained to be

2
2
0
314:16
7 6
0
7 6 0:2178
74
0:1345
0
5
26:5821
0
fd
3
2
Dd
6 Dx 7
7
Dx 0 1 0 0 6
4 DE0 5
q
DE0fd
:
Dd
6 :
6 D x0
6 _
4 DEq
0
DE_
32
3 2
3
Dd
0
0
0
7 6
7
6
0:1086
0 7
76 Dx0 7 6 0 7Dupss
5
5
4
4
DEq
0:5977 0:2
0 5
0
DEfd
3245 100
10000
The electromechanical oscillation modes are as follows:

k2;3 0:0114 j8:2610
In Sect. 2.4.2.1, matrix formed by right eigenvectors is as follows:
v1
V
v2
v3
v4
Thus, the right eigenvectors related to the electromechanical oscillation modes

are as follows:
3
3
2
0:9014
0:9014
7
7
6
6
j0:0237
j0:0237
7
7
6
2 6
v
4 0:0050 + j0:0050 5 v3 4 0:0050 j0:0050 5
0:3868 j0:1930
0:3868 + j0:1930
2
74
Matrix formed by left eigenvectors is as follows:

T w1
W
2
w
3
w
1

w 4 T V
The left eigenvectors related to the electromechanical oscillation modes are as

follows:
T2 0:5554 j0:0001 0:0312 j21:1226 0:1626 j0:1475 0:0003 j0:0003
w
T3 0:5554 j0:0001 0:0312 j21:1226 0:1626 j0:1475 0:0003 j0:0003
w
Thus, according to Eq. (2.64), for the electromechanical oscillation modes, the
residue is calculated to be
2 w
T2 bo cTo
R
v2
0:5554 j0:0001 0:0312 j21:1226 0:1626 j0:1476
3
2
3
2
0:9014
0
7
6
6 0 7
j0:0237
7
6
7
6
6
7
7 0 1 0 0 6
4 0:0050 + j0:0050 5
4 0 5
0:3868 j0:1930
10000
0:0633 j0:0822
0:0003 j0:0003
3 w
T3 bo cTo
R
v3
0:5554 j0:0001 0:0312 j21:1226 0:1626 j0:1475
3
2
3
2
0:9014
0
7
6
6 0 7
j0:0237
7
6
7
6
6
7
7 0 1 0 0 6
4 0:0050 j0:0050 5
4 0 5
0:3868 + j0:1930
10000
0:0633 j0:0822
0:0003 j0:0003
From Eq. (2.104), sensitivity of the electromechanical oscillation modes to the

coefcient of damping torque provided by the PSS can be computed to be
2
22 v22
@k
0:0312 j21:1226j0:0237
w
0:0715

7
@Dpss
M
3
32 v32
@k
0:0312 j21:1226j0:0237
w
0:0715

7
@Dpss
M
2.4 Examples
75
2;3 0:0114 j8:2610, the forward path can be

At the complex frequency k
calculated from Eq. (2.96) as
KA
2;3 T0 1 k2;3 TA K6 KA
K3 k
d0
0:7602 100
2:9885 0:0114 j8:2611 51 0:0114 j8:2611 0:01 0:3245 100

0:8845 j1:1493
2;3 K2
Fpss k
Hence,
2;3
@k
Fpss k2;3 0:0715 0:8845 j1:1493
@Dpss
0:0633 j0:0822
R2;3
Thus, it is conformed that the residue is equal to the forward path of the PSS
multiplied by the sensitivity of electromechanical oscillation modes to the coefcient of damping torque contribution from the PSS.
2.4.4.2
Demonstration by Use of General Linearized Model

of Example Power System
In Sect. 2.4.1.1, state matrix and control vector of state-space model of example
power system are obtained to be
2
Agcdq
0
6 0:045
6
6
6 714:29
6
6 295:81
6
6
6 105:79
6
6
0
6
6
4
0
2
bpss
0
0
6 0 7
6
7
6
7
6 10000 7
6
7
6 0 7
6
7
6
7
6 0 7
6
7
6 0 7
6
7
6
7
4 0 5
0
314:16
0
0
0
0
0:20
0
89:35
318:31
100
0
0
0
0
314:16
0
0:43
1:36
0
0:47
0
0:09
0
0:104
2646
2:88
86:29
3127
3:40
101:98
0
0
1:22
1:88
0:71
3:49
36:0
6116:3 1404:6
6:65
478:4
513:64 5:48
3
0
0:39 7
7
7
1154:5 7
7
135:03 7
7
7
4:50 7
7
7
0
7
7
5
0
46:8
76
Eigenvalues of state matrix are as follows:

1;2 24:12 j971:56
k
3;4 36:32 j359:63
k
k5 35:57
6;7 0:71 j8:44
k
k8 5:71
The pair of electromechanical oscillation modes are as follows:
k6;7 0:71 j8:44
For each of eigenvalues, right eigenvector is calculated to be
2
0:000005 j0:000011
0:000005 j0:000011
6 0:000032 j0:000016 7
6 0:000032 j0:000016 7
6
6
7
7
6
6
7
7
6
6
7
7
0:9507
0:9507
6
6
7
7
6 0:0001 j0:0182 7
6 0:0001 j0:0182 7
6
6
7
7
v1 6
v2 6
7
7
6 0:0369 j0:0014 7
6 0:0369 j0:0014 7
6
6
7
7
6 0:0072 j0:3072 7
6 0:0072 j0:3072 7
6
6
7
7
6
6
7
7
4 0:000569 j0:000026 5
4 0:000569 j0:000026 5
j0:0014
j0:0014
2
2
3
3
3
0:000215 j0:000026
0:000215 j0:000026
0:0527
6 0:000055 j0:000244 7
6 0:000055 j0:000244 7
6 0:006 7
6
6
6
7
7
7
6
6
6
7
7
7
6
6
6 0:0539 7
7
7
0:7251
0:7251
6
6
6
7
7
7
6
6
6 0:1055 7
7
7
0:0163 j0:2196
0:0163 j0:2196
6
6
6
7
7
7
v3 6
v4 6
7
7 v5 6
7
6
6
6 0:2563 7
7
7
0:1654 j0:025
0:1654 j0:025
6
6
6
7
7
7
6 0:0615 j0:6274 7
6 0:0615 j0:6274 7
6 0:4925 7
6
6
6
7
7
7
6
6
6
7
7
7
4 0:0041 j0:0006 5
4 0:0041 j0:0006 5
4 0:0334 5
0:003 j0:0165
0:003 j0:0165
0:8209
2
2
2
3
3
3
0:7004
0:7004
0:000104
6 0:0016 j0:0188 7
6 0:0016 j0:0188 7
6 0:000002 7
6
6
6
7
7
7
6
6
6
7
7
7
6 0:000061 j0:000914 7
6 0:000061 j0:000914 7
6 0:0095 7
6
6
6
7
7
7
6 0:1482 j0:0043 7
6 0:1482 j0:0043 7
6 0:000116 7
6
6
6
7
7
7
v6 6
v7 6
7
7 v8 6
7
6 0:5516 j0:0283 7
6 0:5516 j0:0283 7
6 0:000082 7
6
6
6
7
7
7
6 0:0317 j0:0041 7
6 0:0317 j0:0041 7
6 0:7636 7
6
6
6
7
7
7
6
6
6
7
7
7
4 0:0077 j0:0284 5
4 0:0077 j0:0284 5
4 0:6456 5
2
0:413 j0:0977
0:413 j0:0977
0:000072
2.4 Examples
77
They form the following matrix

v1
V
v2
...
v8
1
; left eigenvectors corresponding to

w
1 w
2 . . . w8 T V
As W
eigenvalues are calculated to be
T
T1
w
T3
w
T6
w
wT8
2
3T
3T
1:2419 j0:4483
1:2419 j0:4483
6 0:0486 j0:262 7
6 0:0486 j0:262 7
6
6
7
7
6 0:4494 j0:0412 7
6 0:4494 j0:0412 7
6
6
7
7
6 0:2153 j3:8659 7 T 6 0:2153 j3:8659 7
6
7 w
7

6
6 1:9686 j0:7712 7 2 6 1:9686 j0:7712 7

6
6
7
7
6 0:0188 j1:3997 7
6 0:0188 j1:3997 7
6
6
7
7
4 0:0288 j1:6542 5
4 0:0288 j1:6542 5
0:6025 j0:5408
0:6025 j0:5408
2
2
2
3T
3T
3T
1:6552 j0:1633
1:6552 j0:1633
0:7411
6 0:1388 j0:9214 7
6 0:1388 j0:9214 7
6 6:8756 7
6
6
6
7
7
7
6 0:1003 j0:0038 7
6 0:1003 j0:0038 7
6 0:000219 7
6
6
6
7
7
7
6 0:2788 j1:9563 7 T 6 0:2788 j1:9563 7 T 6 0:0934 7
6
6
6
7
7
7
4 6
6
7 w
7 w5 6 0:0106 7
6 2:5805 j0:126 7
6 2:5805 j0:126 7
6
7
6 0:0246 j0:1141 7
6 0:0246 j0:1141 7
6 0:000045 7
6
6
6
7
7
7
4 0:0277 j0:1353 5
4 0:0277 j0:1353 5
4 0:00006 5
0:7412 j0:2714
0:7412 j0:2714
1:3311
2
2
3T
3T
0:7422 j0:06
0:7422 j0:06
6 0:2586 j27:632 7
6 0:2586 j27:632 7
6
6
7
7
6 0:000019 j0:00094 7
6 0:000019 j0:00094 7
6
6
7
7
6 0:0034 j0:0146 7 T 6 0:0034 j0:0146 7
6
7 w
7

6
6 0:0004 j0:000131 7 7 6 0:0004 j0:000131 7

6
6
7
7
6 0:000019 j0:000298 7
6 0:000019 j0:000298 7
6
6
7
7
4 0:000139 j0:0002727 5
4 0:000139 j0:0002727 5
0:0499 j0:2027
0:0499 j0:2027
2
3T
0:0803
6 2:7921 7
6
7
6 0:0021 7
6
7
6 0:003 7
7
6
6 0:0283 7
6
7
6 0:000637 7
6
7
4 1:5496 5
0:079
78
According to Eq. (2.64), for the pair of electromechanical oscillation modes, the
residue is calculated to be
6;7 w
T6;7 bo cTo v6;7
R
3
3T 2
2
0
0:7422 j0:06
6 0:2586
j27:632 7 6 0 7
7
7 6
6
7
7 6
6
6 0:000019 j0:00094 7 6 10000 7
7
7 6
6
6 0:0034 j0:0146 7 6 0 7
7
7 6
6
6
7
7 6
6 0:0004 j0:000131 7 6 0 7
7
7 6
6
6 0:000019 j0:000298 7 6 0 7
7
7 6
6
7
7 6
6
4 0:000139 j0:0002727 5 4 0 5
0:0499 j0:2027
0
0 0
0
0:7004
6 0:0016 j0:0188 7
6
7
6
7
6 0:000061
j0:000914 7
6
7
6 0:1482 j0:0043 7
6
7
0 6
7
6 0:5516 j0:0283 7
6
7
6 0:0317 j0:0041 7
6
7
6
7
4 0:0077 j0:0284 5
0:413 j0:0977
0:1772
j0:0114
From Eq. (2.104), the sensitivity of electromechanical oscillation modes to the
coefcient of damping torque provided by the PSS can be computed to be
6;7
62;72 v62;72
@k
0:2586
j27:6320:0016 j0:0188
w

7
@Dpss
M
0:0743
j0:0056
At the complex frequency k6;7 0:71 j8:44, the forward path can be calculated to be
2.4 Examples
79
6;7 aT sI A33 1 bpss3

Fpss k
238
0 1:425
0
3:283 0:617 0:729 2:698

11
3
2
1
C
B
7
6
1
C
B
7
6
C
B
7
6
C
B
7
6
1
C 2
B 0:71 j8:446
7
3
C
B
7
6
10000
1
C
B
7
6
C 6
B
7
6
C 6 0 7
B
5
4
1
7
C 6
B
C 6 0 7
B
7
1
C
B
7
B 2
3C 6
7
B
100 6116:3 1404:6 2646 3127 1154:5 C 6
0
7
6
C 6
B
7
C
B 6 0
7
6:65
478:4
2:88
3:40
135:03 7 C 4 0 5
B 6
B 6
7C
B 6 0
513:64
5:48
86:29 101:98
4:50 7 C
0
7C
B 6
7C
B 6 314:16
0:43
0
1:22
0:71
0
7C
B 6
B 6
7C
@ 4 0
5A
1:36
0
1:88
3:49
0
0
0
2:3823 j0:0253
36:0
46:8
Hence,
R6;7
6;7
@k
Fpss k6;7 0:0743
j0:00562:3823 j0:0253
@Dpss
0:1772
j0:0114
It is thus conrmed that the residue is equal to the forward path of the PSS
multiplied by the sensitivity of oscillation modes to the coefcient of damping
torque contribution from the PSS.
References
1. Demello FP, Concordia C (1969) Concepts of synchronous machine stability as affected by
excitation control. IEEE Trans Power Appar Syst 88(4):316329
2. Heffron WG, Phillips RA (1952) Effect of modern amplidyne voltage regulators on
underexcited operation of large turbine generators. AIEE Trans Power Apparatus Syst
71:692697
3. Demello FP, Laskowski TF (1975) Concepts of power system dynamic stability. IEEE Trans
Power Apparatus Syst 94(3):827833
4. Larsen EV, Swann DA (1981) Applying power system stabilizers part I: general concepts. IEEE
Trans Power Appar Syst 100(6):30173024
Chapter 3
Damping Torque Analysis

of Thyristor-Based FACTS Stabilizers
Installed in Single-Machine Innite-Bus
Power Systems
3.1
A Single-Machine Innite-Bus Power System Installed

with an SVC Stabilizer
3.1.1
Extended HeffronPhillips Model of a Single-Machine

Innite-Bus Power System Installed with an SVC
Stabilizer
3.1.1.1
Nonlinear Mathematical Model of a Single-Machine

Innite-Bus Power System Installed with an SVC Stabilizer
Installation of a shunt thyristor-controlled reactive power compensator, the static

VAR compensator (SVC), is for the voltage support at a key location in a power
system. An additional stabilizing signal can be superimposed on the voltage control
loop of the SVC to provide extra damping to the power system. This supplementary
damping controller is named the SVC stabilizer in this book. Figure 3.1 shows the
conguration of a thyristor-controlled reactor and xed capacitor (TCR-FC) type of
SVC with an additional damping control loop, i.e., the SVC stabilizer.
Figure 3.2 shows the conguration of a single-machine innite-bus power
system installed with an SVC, where bsvc is the equivalent admittance of the SVC.
For the TCR-FC, SVC [1]
jbsvc 1 Ca=jxsvcl 1=jxsvcc
3:1
2a
1 and a is the ring angle.
where Ca 2asin
2p
Isb Its Is Its jbsvc V

s
s jxsbIsb V
s V
b jxsbIts xsb bsvc V
b
V
DOI 10.1007/978-1-4899-7696-3_3
3:2
81
3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers
82
x svcl
Vs
Vsref
voltage
controller
x svcc
firing circuit
/2
ysvc
SVC
stabilizer
Fig. 3.1 Conguration of a TCP-FC SVC with damping control function
Fig. 3.2 A single-machine

installed with an SVC
Vt
Vs
xts
xsb
I sb
Its
Is
t jxtsIts V
s
V
Vb
jb svc
3:3

s jxsb Its Vb
V
csvc
3:4
where csvc 1 xsb bsvc :

By substituting Eq. (3.4) into Eq. (3.3) it can have
t jxlRIts V
b =csvc
V
3:5
where xlR xts xsb =csvc .

Equation (3.5) indicates that the power system installed with the SVC is
equivalent to a power system without the SVC installed. The equivalent system is

of a line impedance xlR connected to an innite bus with a voltage being cVsvcb .
83
Therefore, the mathematical model of the single-machine innite-bus power system

installed with the SVC as shown in Fig. 3.2 can be obtained simply by modifying
the model of the system without the SVC of Eqs. (2.37) and (2.38). The modication is to replace xt and Vb in Eqs. (2.37) and (2.38) by xlR and cVsvcb , respectively,
to obtain that
:
d xo x 1
1
:
x Pm Pt D(x 1
M
1
0
E_ q 0 Eq Efd
Tdo
1
KA
0
E_ fd E0fd
Vtref Vt
TA
TA
Pt
E0q Vb
sin d
V2b xq x0d
sin 2d
c2svc 2 x0dR xqR
csvc x0dR
E0q xdR xd x0d Vb cos d
Eq 0
csvc x0dR
xdR
Efd
vtd
3:6
3:7
Efd0 E0fd
q
xt E0q Vb x0d cos d
xq Vb sin d
; vtq 0
; Vt v2td + v2tq
0
csvc xdR
csvc xqR
xdR
where x0dR x0d xlR ; xdR xd xlR ; xqR xq xlR :

Similar modication of Eq. (2.33) gives
itsq
E0q
Vb cos d
csvc x0dR
Vb sin d
csvc xqR
itsd
x0dR
3:8
where itsd and itsq are the d and q component of line current Its , respectively.
3.1.1.2
Extended HeffronPhillips Model
By linearizing the rst two equations in Eq. (3.7), it can have

DPt K1 Dd K2 DE0q Kp Dbsvc
DEq K3 DE0q K4 Dd Kq Dbsvc
3:9
84
where
K1
E0q0 Vb
cos d0
csvc0 x0dR
Vb
K2
sin d0
csvc0 x0dR
V2b xq x0d
cos 2d0
c2svc0 x0dR xqR

0
V2b xq x0d sin(2d0 xsb xts x0d
@Pt Vb Eq0 sin d0 xsb xts x0d
Kp

@bsvc 0
2c3svc0 x0dR 2 xqR
csvc0 x0dR 2

V2 xq x0d sin(2d0 xsb xts xq
b
2c3svc0 x0dR xqR 2

xsb xts x0d Vb E0q0 sin d0 V2b xq x0d sin(2d0

csvc0 x0dR
csvc0 x0dR
2c2svc0 x0dR xqR

2
Vb xq x0d sin(2d0 xsb xts xq

csvc0 xqR
2c2svc0 x0dR xqR
xdR
x0dR
xd x0d Vb sin d0
K4
csvc0 x0dR
E0q0 xdR x2
E0q0 x2

xd x0d Vb cos d0
Kq 0 2 2sb 0 2sb
xsb xts x0d
2
0
x
xdR csvc0
csvc0 xdR
dR csvc0
K3
In deriving Kp and Kq above, following equations are used

csvc x0dR csvc x0d xts xsb =csvc csvc x0d xts xsb
1 xsb bsvc x0d xts xsb
3:10
csvc xq 1 xsb bsvc xq xts xsb

By linearizing the last three equations in Eq. (3.7), it can have
xq Vb cos d0
xq Vb sin d0
Dd
xsb xsb xq Dbsvc
csvc0 xqR
csvc0 xqR 2
Vb x0d sin d0
xLR
Dvtq
Dd 0 DE0q
3:11
csvc0 x0dR
xdR
"
#
xlR E0q0 x2
E0q0 x2
Vb x0d cos d0
0 2 2sb 0 2sb
xsb xsb x0d Dbsvc
xdR csvc0
xdR csvc0
csvc0 x0dR 2
Dvtd
85
Hence,
DVt
vtq0
vtd0
Dvtd
Dvtq K5 Dd K6 DE0q Kv Dbsvc
Vt0
Vt0
3:12
where
vtd0 xq Vb cos d0 vtq0 Vb x0d sin d0

Vt0 csvc0 xqR
Vt0 csvc0 x0dR
vtq0 xLR
K6
Vt0 x0dR
vtd0 xq Vb sin d0
Kv
xsb xsb xq
Vt0 csvc0 xqR 2
"
#
xlR E0q0 x2sb
E0q0 x2sb
vtq0
Vb x0d cos d0
0
0 2 2 0 2
x x xd
0 2 sb sb
Vt0
xdR csvc0
xdR csvc0
csvc0 xdR
K5
By substituting Eqs. (3.9) and (3.12) into the linearized equation of Eq. (3.6),
it can have
Dd_ xo Dx
1
Dx_ K1 Dd DDx K2 DE0q Kp Dbsvc
M
1
0
_
DEq 0 K3 DE0q K4 Dd Kq Dbsvc DEfd
Td0
1
KA
0
0
DE_ fd DE_ fd
K5 Dd K6 DE0q Kv Dbsvc
TA
TA
3:13
Equation (3.13) is the extended HeffronPhillips model of the single-machine

innite-bus power system installed with the SVC, which is shown in Fig. 3.3.
3.1.1.3
Extended HeffronPhillips Model with Both the SVC Voltage

and Damping Control Function Included
By substituting Its itsd jitsq into Eq. (3.4), it can have

Vb sin d itsq xsb
csvc
Vb cos d itsd xsb
vsq
csvc
vsd
3:14
86
K1
0
s
1
D + sM
bsvc
K2
K4
Kp
Kq
K5
Kv
SVC
1
K 3 + sTd0 '
- -
KA
1 + sTA
K6
Fig. 3.3 Extended HeffronPhillips model of the single-machine innite-bus power system
installed with an SVC stabilizer
s , respectively. By using
where vsd and vsq are the d and q component of voltage V
Eq. (3.8), from Eq. (3.14), it can be obtained that

xsb Vb sin d
Vb sin d
csvc
csvc xqR
csvc

0
xsb Eq Vb cos d
1
vsq
Vb cos d 0
xsb
csvc
csvc x0dR
xdR
q
Vs v2sd v2sq
vsd
b sin d
Vb sin d V
csvc xqR xsb
3:15
Linearization of Eq. (3.15) is as follows:

DVs C1 Dd C2 DE0q C3 Dbsvc
3:16
where

vsq0 @vsq
@Vs vsd0 @vsd
C1
@d 0 Vs0 @d 0 Vs0 @d 0

vsq0 1
vsd0 1
xsb Vb cos d0
Vb sin d0
Vb0 cos d0
Vb0 sin d0
x
sb
Vs0 csvc0
csvc0 xqR
Vs0 csvc0
csvc0 x0dR
87

vsq0 @vsq vsq0 1 xsb
@Vs vsd0 @vsd
C2 0

@Eq
Vs0 @E0q
Vs0 @E0q
Vs0 csvc0 x0dR
0
0
0

vsq0 @vsq
@Vs vsd0 @vsd
C3
@bsvc 0 Vs0 @bsvc 0 Vs0 @bsvc 0

"
#
vsd0 xsb vsd0
x2sb Vb sin d0

xq xts
Vs0 csvc0
csvc0 csvc0 xqR 2

xsb E0q0 Vb cos d0
vsq0 xsb
V
cos
d

x
b
0
sb
Vs0 c2svc0
x0dR
csvc0 x0dR

E0q0 x3
1
Vb cos d0
0 2 2sb 2 0 2 x2sb x0d xts

csvc0
xdR csvc0
csvc0 xdR
For the simplicity of discussion, it is assumed that both the SVC voltage and
damping controller adopt the proportional control law with a gain CV and CS ,
respectively. If the feedback signal of the SVC stabilizer is the rotor speed of
generator, from Fig. 3.1, it can have
Da CV DVs CS Dx
3:17

bsvc
1
1 C(a

xsvcc
xsvcl
3:18
Hence,
_ 0
_ 0
Ca
Ca
Da
CV DVs CS Dx
xsvcl
xsvcl
4
_ 0 2 2cos 2a0 [ 0
[ Ca
where
p
p
Dbsvc
3:19
By substituting Eq. (3.16) into Eq. (3.19), the following equation can be
obtained
:
Dbsvc
Ca0
CV C1 Dd CV C2 DE0q CV C3 Dbsvc CS Dx

Xsvcl
3:20
Thus,
Dbsvc
CV C1 Dd CV C2 DE0q CS Dx
x
: svcl CV C3
Ca0
3:21
88
K1
0
s
1
D + sM
SVC
Cs
C + C3 C V
K2
C 2 C V
C + C3 C V
C1CV
C + C3 C V
bsvc
K4
Kp
1
K 3 + sTd0 '
Kq
K5
Kv
- +
KA
1 + sTA
K6
Fig. 3.4 Extended PhilipHeffron model with SVC voltage and damping control function
included
By taking account of Eq. (3.21), Fig. 3.3 can be further extended to include the
SVC voltage and damping control function as shown in Fig. 3.4, where C :xsvcl .
Ca0
3.1.1.4
Calculation of Initial Compensation
This section demonstrates how the initial compensation of the SVC installed in the
single-machine innite-bus power system shown in Fig. 3.2 can be calculated to
89
satisfy the requirement to maintain the voltage prole of the system. The initial
conditions to compute the initial compensation, bsvc0 , of the SVC are the magnitude
of busbar voltage, Vt0 ; Vb0 ; Vs0 , and the active power supplied by the generator, Pt0 .
Figure 3.5 shows the circuit model of the system as shown in Fig. 3.2. From
Fig. 3.5, it can be obtained that
s0 V
b
V
Psb0 jQsb0
V
s0
jxsb
3:22
Since the resistance of the transmission lines is ignored, Psb0 Pt0 . By choosing
s0 as the reference phasor on x-axis, it can have V
s0 V
Vs0 . From
V
s0
Eq. (3.14), the following equation can be obtained
b Vs0 xsb Qsb0 jPsb0 Vs0 xsb Qsb0 j xsb Psb0
V
Vs0
Vs0
Vb
The above equation gives
V2b

2
2
xsb
xsb
Vs0
Q
Psb0
Vs0 sb0
Vs0
3:23
Solution of Eq. (3.23) is as follows
Qsb0
Vt0
Pt 0 + jQ t 0
2
3
s

2
Vs0 4
x
sb
Vs0 V2b
Psb0 5
xsb
Vs0
jx ts
Pts0 + jQ ts0
I ts0
V s0
Psb0 + jQsb0
Ps0 + jQs0
bsvc0
j
jx sb
3:24
Pb0 + jQ b0
Vb
Isb0
Is
Fig. 3.5 Circuit model of the single-machine innite-bus power system installed with SVC
90
Similarly, from the following equation

t0 V
s0
V
Pts0 jQts0
V
s0
jxts
it can have
V2t0

2
2
xts
xts
Vs0
Q
Pts0
Vs0 ts0
Vs0
3:25
Qts0
2s
3

2
Vs0 4
x
ts
V2t0
Pts0 Vs0 5
xts
Vs0
3:26
Hence,
Ps0 jQs0 0:0 j(Qts0 Qsb0
Since
s0 jbsvc0 Is0
V
it can have
jbsvc0 Is0 V
Ps0 jQs0
s0 V
V
s0
s0
Finally from the above equation, the initial compensation is calculated to be
bsvc0
Ps0 jQs0
Q
Q Q
s0
sb0 2 ts0
V2s0
Vs0
jV2s0
3.1.2
Damping Torque Analysis of SVC Stabilizer
3.1.2.1
Electric Torque Provided by the SVC Stabilizer
3:27
From Fig. 3.4, the forward path of the SVC stabilizer to the electromechanical
oscillation loop of generator can be established as shown in Fig. 3.6.
From Fig. 3.6, it can be seen that the SVC stabilizer contributes the electric
torque to the electromechanical oscillation loop through two paths. The electric
91
Cs
C + C3 C V
K2
C 2 C V
C + C3 C V
bsvc
Kp
Kq
Kv
1
K 3 + sTd0 '
KA
1 + sTA
K6
Fig. 3.6 Forward path of the SVC stabilizer
torque through the block, KP , is called the direct electric torque, denoted by DTed .
That through the blocks, Kq and KV , is named the indirect electric torque, denoted
by DTei . According to the damping torque analysis introduced in Chap. 2, electric
torque can be decomposed into a damping torque and a synchronizing torque at the
angular oscillation frequency, xs , that is
DTed DTsd DTdd Tsd Dd Tdd Dx
DTei DTsi DTdi Tsi Dd Tdi Dx
3:28
where Tsd ; Tsi ; Tdd ; Tdi are direct, indirect synchronizing torque and direct, indirect
damping torque coefcients, respectively. From Fig. 3.6, the transfer function of
forward path of the SVC stabilizer to the electromechanical oscillation loop of
generator can be derived to be
92
KA
1
K3 sT0d0 1 sTA
Fsvc s) Kp K2 Kq
K2 Kv
KA
KA
1
1
1 K6
1 K6
K3 sT0d0 1 sTA
K3 sT0d0 1 sTA
K2 Kq 1 sTA Kv KA
Kp
K3 sT0d0 1 + sTA + K6 KA
1
K3 sT0d0
3:29
Hence, the electric torque provided by the SVC stabilizer is as follows
DTet Fsvc jxs Dbsvc
3:30
From Fig. 3.6, it can be seen that the signals are attenuated by the exciter and the
AVR before they form the indirect damping toque. Hence, normally Tdd Tdi .
From Fig. 3.6 and Eqs. (3.29) and (3.30), it can be seen that the coefcient Kp is the
dominant weight parameter in determining the amount of damping torque contribution from the SVC stabilizer. Hence, in the following section, how the effectiveness of the SVC stabilizer is affected by various factors is examined by
checking the dominant weight parameter KP .
3.1.2.2
Damping Control as Affected by the Load Conditions

Pt0
E0q0 Vb
csvc0 x0dR
sin d0
V2b xq x0d
sin 2d0 Pt10 Pt20
c2svc0 2 x0dR xqR
3:31
E0q0 Vb
V2b xq x0d
where Pt10 csvc0
x0dR sin d0 ; Pt20 c2svc0 2 x0dR xqR sin 2d0 :
From Eq. (3.9) and by using Eq. (3.31), the weight parameter Kp can be obtained
to be

xsb xts x0d Vb E0q0 sin d0 V2b xq x0d sin(2d0
Kp

csvc0 x0dR
csvc0 x0dR
2c2svc0 x0dR xqR

2
0
Vb xq xd sin(2d0 xsb xts xq

csvc0 xqR
2c2svc0 x0dR xqR

0
xts xq
xsb xts xd
Pt0
Pt20
csvc0
x0dR
xqR
3:32
From Fig. 3.7 and Eq. (3.32), it can be seen that at a higher load condition,
difference between Pt0 and Pt20 is bigger. Thus, the higher the load condition is, the
bigger Kp is, the more damping torque is provided by the SVC stabilizer. Hence,
effectiveness of the SVC stabilizer should increase with system load conditions.

Fig. 3.7 Pt d curve of the
power system
93
Pt 0 = Pt10 Pt 20
Pt10
Pt 20
/2
/4
3.1.2.3
Influence of Parameters of the Generator
Since xq [ x0d , it should have

xts x0d
xts xq
1
1
x0dR
xqR
1 c xxsb x0
1 c xxsb x
svc0 ts
svc0 ts
q
d
3:33
Equations (3.32) and (3.33) indicate that Kp is affected not only by the load
conditions of the power system, but also by the parameters of the generator, x0d and
xq , such that:
1. If the difference between the values of x0d and xq is small (for example, for the
generator in a hydropower station, value of x0d is around 0.20.35 and that of xq
is around 0.450.7) or if the system operates at a higher load condition
xts x0
xts xq
(Pt0 Pt20 ), x0 d Pt0 [ xqR Pt20 such that Kp [ 0, the SVC stabilizer
dR
will provide the power system with positive damping.
xts x0
xts xq
2. If xq is much greater than x0d such that x0 d xqR ; (for example, for a
dR
generator in a thermal power plant, x0d is around 0.15 to 0.24 and xq is around
94
1.22.2) and when the system operates at a lower load condition such that the
difference between Pt0 and Pt20 is small, there is a possibility that
xts x0d
xts xq
x0dR Pt0 \ xqR Pt20 such that Kp \0. This means that the SVC stabilizer
which supplies positive damping torque at a higher load condition may provide
the power system with negative damping torque at a lower load condition.
xts x0
xts xq
3. Around the operating point where x0 d Pt0 xqR Pt20 such that Kp 0,
dR
the damping torque provided by the SVC stabilizer is zero. At this operating
point, the SVC stabilizer will lose its capability to suppress power system
oscillation. This load condition is called the dead point of the SVC stabilizers
function.
3.1.2.4
Electric Length of the Transmission Line
Dene z to be an index of the length of the transmission line and take z0 1. The
reactance of the transmission line is proportional to its electric length, i.e., the index z.
Kp can be written as a function z as (see Eq. (3.32))

zxts x0d
zxts xq
zxsb
Pt0
Pt20
Kp z)
xq z(xts xsb =csvc0
csvc0 x0d z(xts xsb =csvc0
3:34
For the simplicity of discussion, it is assumed that at the steady-state operation,

bsvc0 0 such that csvc0 1. Hence,

zxts x0d
zxts xq
Pt0
Pt20
x0d z(xts xsb
xq z(xts xsb

Kp z) zxsb
It can have

zxts x0d
zxts xq
@
Kp z) xsb 0
Pt0
Pt20
@z
xd z(xts xsb
xq z(xts xsb
(

zxts x0d xts xsb
xts
Pt0
Pt0
zxsb 0
xd z(xts xsb
x0d z(xts xsb 2
)

zxts xq xts xsb
xts

Pt20
Pt20
xq zxts xsb
xq z(xts xsb 2
3:34
95
Hence,

xts x0d
xts xq
@Kp z)
P
P
x

sb
to
t02
@z zz0 1
x0d xts xsb
xq xts xsb
(

xts x0d xts xsb
xts
Pt0 0
xsb 0
Pt20
xd xts xsb
xd xts xsb 2
)

xts xq xts xsb
xts

Pt20
Pt20
xq xts xsb
xq xts xsb 2

xsb xts x0d
xsb xts x0d xts xsb
xsb xts
0
Pt0 0
Pt0
Pt0
xd xts xsb
xd xts xsb
x0d xts xsb 2

xsb xts xq
xsb xts xq xts xsb
xsb xts
Pt20
Pt20

Pt20
xq xts xsb
xq xts xsb
xq xts xsb 2

xsb xts x0d
xsb xts x0d xts xsb
xsb xts
Pt0 0
Pt0
0
Pt0
xd xts xsb
xd xts xsb
x0d xts xsb 2

xsb xts xq
xsb xts xq xts xsb
xsb xts
Pt20
Pt20

Pt20
xq xts xsb
xq xts xsb
xq xts xsb 2

xsb xts x0d x0d xsb xts x0d xts xsb
Pt0
x0d xts xsb 2

xsb xts xq xq xsb xts xq xts xsb

Pt20

2
xq xts xsb

xsb x0 xts x0d
xsb xq xts xq
xsb xts
xsb xts
0 d
P
P

P
Pt20

t0
t0
0
2 t20
xd xts xsb
x
xts xsb
xd xts xsb 2
q
xq xts xsb

xsb x0d xts x0d
xsb xq xts xq
xsb xts
xsb xts
Pt0
Pt20 0 Pt0
Pt20
xdR
xqR
x02
x2qR
dR

xsb x0d xts x0d
xsb xq xts xq
Pt0 Pt20
Pt0
Pt20 xsb xts 0
02
2
xdR xqR
xdR
xqR
3:35
Thus,

xsb x0d xts x0d xsb xts x0dR
xsb xq xts xq xsb xts xqR
@Kp z)
P

Pt20
t0
@z zz0 1
X02
X2qR
dR

2

2
xsb xts x0d xts x2sb
xsb xts xq xts x2sb
Pt0
Pt20
x02
x2qR
dR
3:36
When Kp [ 0

xts x0d
xts xq
Pt0 [
Pt20
x0dR
xqR
3:37
96
Hence, normally it should have

2
2

xts x0d
xts xq
Pt0 [
Pt20
x0dR
xqR
3:38
Using Eqs. (3.36) and (3.38), from Eq. (3.35), it can be obtained that
@Kp z)
[0
@z
3:39
Equation (3.39) means that for the SVC stabilizer designed to supply positive
damping to power system oscillations, it will be more effective when the electric
length of transmission lines increases.
3.1.2.5
Installing Location of the SVC
If it is assumed that the total length of the transmission line is xed, that is,
xts xsb xt , from Eq. (3.32), it can be obtained that (assuming csvc0 1)

Kp xsb

xsb xt xsb x0d
xsb xt xsb xq
P

P
t0
t20
xt x0d
xq xt
3:40
Dene the following two functions of xsb

xsb xt xsb x0d
g(xsb
Pt0
x0 xt
d

xsb xt xsb xq
h(xsb
Pt20
xq xt
3:41
Kp xsb g(xsb h(xsb
3:42
Thus,
By solving the following two equations

@
xt 2xsb x0d
g(xsb
Pt0 0
@xsb
x0d xt
xt 2xsb xq
@
h(xsb
Pt20 0
@xsb
xq xt
x0d xt
3:43
x x
xsb1 and xsb q 2 t xsb2 . Hence, g(xsb and
0
x x
x x
h(xsb achieve its maximum value at xsb d 2 t xsb1 and xsb q 2 t xsb2 ,
respectively.
It can have xsb
97
K p (x sb )
g(x sb )
K p (x sb )
h(x sb )
x sb
0
xt
2
x sb1
x sb2
xt
Fig. 3.8 Illustration on the most effective installing location of the SVC stabilizer
From Eqs. (3.42) and (3.43), the curve of Kp xsb can be sketched as shown in
Fig. 3.8 (Pt0 Pt20 ). Figure 3.8 demonstrates that when xsb x2t , Kp xsb achieves
its maximum value. This means that when the SVC is installed near the middle
point of the transmission line, the SVC stabilizer contributes the most amount of
damping torque and hence is most effective in suppressing power system
oscillations.
3.2

with a TCSC or TCPS Stabilizer
3.2.1

Innite-Bus Power System Installed with a TCSC
or TCPS Stabilizer
3.2.1.1

Innite-Bus Power System Installed with a TCSC
It was discovered decades ago that power oscillations along transmission lines can
be effectively suppressed by switching series capacitors in and out from the lines
after a fault [2, 3]. Nowadays, the high-speed thyristor-controlled series compensator
98
(TCSC) has made it practically possible. A TCSC can be simply represented as a

capacitor whose capacitance is adjustable within a certain range as shown in Fig. 3.9,
which is a single-machine innite-bus power system installed with a TCSC.
Obviously, installation of the TCSC only changes the equivalent reactance of the
transmission line connecting the generator and the innite busbar. Therefore, the
mathematical model of the system can be established by simply modifying Eqs. (2.
37) and (2.38). The modication is just to change the expressions of the equivalent
reactance of the transmission line as follows
d_ xo x 1
1
x_ Pm Pt D(x 1
M
1
0
E_ q 0 Eq Efd
Tdo
1
KA
0
E_ fd E0fd
Vtref Vt
TA
TA

E0q Vb
V2b xq x0d
Pt 0 sin d
sin 2d
xdR
2 x0dR xqR

E0q xdR
xd x0d Vb cos d
Eq 0
x0dR
xdR
Efd
vtd
3:44
3:45
Efd0 E0fd
q
xt E0q Vb x0d cos d
xq Vb sin d
; vtq 0
;
V
v2td v2tq
t
x0dR
xqR
xdR
where
xdR xd xL xtcsc ; x0dR x0d xt xtcsc
xqR xq xL xtcsc
Vt
3:46
Vb
jx tcsc
jx t
Fig. 3.9 A single-machine innite-bus power system installed with a TCSC
99
By linearizing Eq. (3.45), it can be obtained that

DPt K1 Dd K2 DE0q Kp Dxtcsc
DEq K3 DE0q K4 Dd Kq Dxtcsc
DVt
K5 Dd K6 DE0q
3:47
Kv Dxtcsc
where
K1
E0q0 Vb

V2b xq x0d
cos d0
cos 2d0
x0dR xqR
x0dR
Vb
K2 0 sin d0
xdR

0

@Pt Vb Eq0 sin d0 V2b xq x0d sin 2d0 0
Kp

xdR xqR

02
02
2
@x
x
2x x
tcsc 0
K3
K4
Kq
K5
K6
Kv
dR
dR qR
xdR
0
xdR

xd x0d Vb sin d0
x0dR

0
xd x0d Vb cos d0
@Eq Eq0 xdR
02
@xtcsc 0
xdR
x02
dR
0
vtd0 xq Vb cos d0 vtq0 Vb xd sin d0

Vt0
xqR
Vt0
x0dR
vtq0 xLR
Vt0 x0dR

@Vt vtd0 xq Vb sin d0 vtq0
xL
1
Vb x0d cos d0
0
0 Eq0
xdR
@xtcsc 0 Vt0
Vt0
x2qR
x02
x02
dR
dR
By substituting Eq. (3.46) into the linearized equation of Eq. (3.44), it can have
:
D d xo Dx
1
:
D x K1 Dd DDx K2 DE0q Kp Dxtcsc
M
1
0
DE_ q 0 K3 DE0q K4 Dd Kq Dxtcsc DE0fd
Td0
1
KA
DE0fd DE0fd
K5 Dd K6 DE0q Kv Dxtcsc
TA
TA
3:48
100
K1
0
s
1
D + sM
x t csc
K2
K4
Kp
Kq
K5
Kv
TCSC
1
K 3 + sTd0 '
- +
KA
1 + sTA
K6
installed with a TCSC stabilizer

innite-bus power system installed with the TCSC stabilizer, which is shown in
Fig. 3.10.
3.2.1.2

Innite-Bus Power System Installed with a TCPS
The phase shifters have been used to regulate the steady-state power flow in power
systems. The idea of applying online control of phase shifters was proposed decades ago. However, the low-speed mechanical tap changers precluded the use of
online control of phase shifters to improve power system dynamic performance.
With the advances in power electronics, high-power high-speed electronic switches
make it possible to realize real-time control of a phase shifter. The feasibility of a
thyristor-controlled phase shifter (TCPS) as a means to adjust the value of
quadrature component of line voltage has been well recognized.
Figure 3.11 illustrates the arrangement of a TCPS installed along a transmission
line. Its function can be simply expressed as a phase shifting of line voltage as
shown in Fig. 3.11. Hence, for a single-machine innite-bus power system installed
with a TCPS as shown in Fig. 3.12, system dynamic Eqs. (2.37) and (2.38) need to
be modied only in the phase relationship as follows.
101
V'
Vector diagram
Symbol
V
V'
V
Converter
V'
TCPS
Fig. 3.11 Arrangement of a TCPS and its simple representation
Vt +
Vt
TCPS
Vb
jx t
Fig. 3.12 The single-machine innite-bus power system installed with a TCPS
:
d xo x 1
1
:
x Pm Pt D(x 1
M
_E0 1 Eq Efd
q
T0do
1
KA
0
E_ fd E0fd
Vtref Vt
TA
TA
Pt
E0q Vb
sin(d /
3:49
V2b xq x0d
sin2(d /
2 x0dR xqR
x0dR
E0q xdR xd x0d Vb cos(d /
Eq 0
x0dR
xdR
xq Vb sin(d /
vtd
xqR
xL E0q Vb x0d cos(d /
vtq 0
x0dR
xdR
3:50
102
By linearizing Eq. (3.50), it can have

DPt K1 Dd K2 DE0q Kp D/
DEq K3 DE0q K4 Dd Kq D/
DVt
K5 Dd K6 DE0q
3:51
Kv D/
where
K1
E0q0 Vb
cos d0
V2b xq x0d
cos 2d0
x0dR xqR
x0dR
Vb
K2 0 sin d0
xdR

0
V2 xq x0d
@Pt Eq0 Vb
Kp
cos d0 b 0
cos 2d0

0
x x
@/
x
0
dR
dR qR
xdR
K3 0
xdR
xd x0d Vb sin d0
K4
x0dR

@Eq xd x0d Vb sin d0
Kq
@/ 0
x0dR
vtd0 xq Vb cos d0 vtq0 Vb x0d sin d0
K5

Vt0
xqR
Vt0
x0dR
vtq0 xLR
K6
Vt0 x0dR

@Vt vtd0 xq Vb cos d0 vtq0 Vb x0d sin d0
Kv

@/ 0 Vt0
xqR
Vt0
x0dR
By substituting Eq. (3.51) into the linearized equation of Eq. (3.49), it can be
obtained that
:
D d xo Dx
1
:
D x K1 Dd DDx K2 DE0q Kp D/
M
1
0
DE_ q 0 K3 DE0q K4 Dd Kq D/
Td0
1
KA
DE0fd DE0fd
K5 Dd K6 DE0q Kv D/
TA
TA
3:52

innite-bus power system installed with the TCSC stabilizer, which is shown in
Fig. 3.13.
103
K1
0
s
1
D + sM
K2
K4
Kp
Kq
K5
Kv
TCPS
1
K 3 + sTd0 '
- +
KA
1 + sTA
K6
installed with a TCPS stabilizer
3.2.2
Damping Torque Analysis of TCSC and TCPS

Stabilizers
3.2.2.1
General Expression of Damping Torque Contributed

by the TCSC and TCPS Stabilizers
For the simplicity of discussion, it is assumed that both the TCSC and TCPS
stabilizers adopt the deviation of rotor speed of the generator as the feedback signal.
Denote the transfer function of the stabilizers by Ktc Ttc s). The output control signal
of the TCSC and TCPS stabilizers is as follows
Dutc Ktc Ttc s)Dx
3:53
where Dutc is Dxtcsc or D/ for the TCSC stabilizer or TCPS stabilizer, respectively.
From Figs. 3.10 and 3.13, the forward path of the stabilizer can be obtained and
shown in Fig. 3.14.
From Fig. 3.14, it can be seen that the direct damping torque contribution, DTdd ,
is through Kp . It should be much greater than the indirect damping torque, DTdi ,
104
Td
Tdi
Tdd
u tc
K2
Kp
Kq
Kv
1
K 3 + sTd0 '
KA
1 + sTA
K6
Fig. 3.14 Forward path of the TCSC or TCPS stabilizer
because Dutc is attenuated by two rst-order lag blocks before it forms the indirect
damping torque. That is
DTd Tdd Dx Re[Kp Ktc Ttc jxs Dx
3:54
From Eq. (3.54), it can be seen that coefcient Kp weighs the amount of damping
torque contributed by the TCSC and TCPS stabilizers to the power system.
3.2.2.2
Damping Torque Provided by the TCSC Stabilizer [4]

E0q0 Vb
V2b xq x0d
Pt0 0 sin d0
sin 2d0 Pt10 Pt20
xdR
2 x0dR xqR
3:55
where
Pt10
Pt20
E0q0 Vb
x0dR
sin d0
V2 xq x0d
b 0
sin 2d0
2 xdR xqR
3:56
105
Hence, for the TCSC stabilizer, from Eq. (3.47), it can be obtained that

0

@Pt Vb Eq0 sin d0 V2b xq x0d sin2d0 0

xdR xqR
Kp

02
2
02
@xtcsc 0
xdR
2xdR XqR

0
2
0
1 Vb Eq0 sin d0 Vb xq xd sin(2d0
1 V2b xq x0d sin2d0
0

2x0dR xqR
xdR
xqR
2x0dR xqR
x0dR
1
1
0 Pt10
Pt20
xdR
xqR
3:57
Since xqR [ x0dR and Pt0 [ Pt20 as shown by Fig. 3.15,
Kp [ 0
3:58
Equation (3.58) indicates that when the load conditions of the power system
change, a properly designed TCSC stabilizer can always provide positive damping
torque to the power system.
At a higher load condition, difference between Pt0 and Pt20 is greater as shown
in Fig. 3.15 and hence Kp is bigger as shown in Eq. (3.57). This means that
Fig. 3.15 Pt d curve of the

power system installed with a
TCSC
Pt 0 = Pt10 Pt 20
Pt10
Pt 20
/2
/4
106
effectiveness of the TCSC stabilizer in damping power system oscillations increases

with the increase in system load conditions.
!
@Kp
Pt0 Pt20
02 2 \0
@xt
xdR xqR
3:59
Equation (3.59) indicates that the TCSC stabilizer is less effective when the
equivalent reactance of transmission line increases if the load condition keeps
unchanged.
3.2.2.3
Damping Torque Provided by the TCPS Stabilizer [5]
For the TCPS stabilizer, from Eq. (3.51), it can have

@Pt
Kp
Kp1 Kp2
@/ 0
3:60
where
Vb E0q0 cos d0
x0
dR 0
2
V xq xd cos 2d0
b
2x0dR xqR
Kp1
Kp2
3:61
Kp1 and Kp2 are shown in Fig. 3.16. From Fig. 3.16, it can be seen that Kp [ 0.
Hence, when the load conditions of the power system change, a properly designed
TCPS stabilizer can always supply positive damping torque to the power system.
!
@Kp
Kp1 Kp2
02 2 \0
@xt
xdR xqR
3:62
Equation (3.62) indicates that the TCPS stabilizer is less effective when the
equivalent reactance of the transmission line increases if the loading condition is
unchanged.
3.3 An Example Power System Installed with an SVC Stabilizer
107
Pt 0
a
K p1
K p2
/2
/4
a=
Vb E q0 'sin 0
x d '
b=
Vb 2 (x q x d ') sin 2 0
2x d ' x q
Fig. 3.16 Examination of the damping effectiveness of the TCPS stabilizer
3.3
An Example Power System Installed with an SVC

Stabilizer
Parameters of an example single-machine innite-bus power system with an SVC

installed in Fig. 3.2 are (in p.u. except indicated)
0
Generator: xd 1:0; xq 0:8; x0d 0:15; M 6:0 s:; D 0; Td0
5:044 s:
The AVR: KA 20:0; TA 0:01 s:
Transmission line: Xts 0:3; Xsb 0:3:
The SVC: xsvcl 1; xsvcc 1; Kvp 1; Kvi 8:
Steady-state operating condition: Pt0 0:5; Vt0 1:0; Vs0 1:0; Vb 1:0
108
3.3.1
Linearized Model of Example Power System
3.3.1.1
Linearized Model with the Transfer Function of the SVC

Voltage Controller Included
The SVC is equipped with a proportionalintegral (PI) voltage controller. The added
damping controller adopts the locally available deviation of line active power as the
feedback signal and the transfer function of a conventional PSS. That is (see Fig. 3.1)

Kvi
a a0 Kvp
Vsref Vs usvcs
s
3:63
where usvc-s is the output stabilizing signal of the SVC stabilizer, i.e.,
usvcs Ks
1 sT2 1 sT4
Ptref Pt
1 sT1 1 sT3
Linearization of Eq. (3.63) is as follows

Kvi
Da Kvp
DVs Dusvcs
s
3:64
0
Da, the above equation can be converted to
From Eq. (3.19) with Dbsvc Cxa
svcl
Dbsvc
:
:

Ca0
Kvi
Ca0

Kvp
Dusvcs
DVs
xsvcl
xsvcl
s
3:65
Substituting Eq. (3.16) into the above equation, it can have

Dbsvc
:

i C: a
Ca0
Kvi h
0
Kvp
Dusvcs
C1 Dd C2 DE0q C3 Dbsvc
xsvcl
xsvcl
s
3:66
The above equation can give
Dbsvc Fsvc1 s)(C1 Dd C2 DE0q Fsvc2 s)Dusvcs
3:67
where
:

:
Kvi
0
Cxa
K
vp
s
1
Ca0
svcl
Fsvc1 s)
:
:

; Fsvc2 s)

Ca0 K Kvi xsvcl
Kvi
0
1
C
K
1 C3 Cxa
vp
3
vp
s
xsvcl
s
svcl
109
K1
0
s
1
D + sM
SVC
Fsvc2 (s)
K2
C2 Fsvc1 (s)
u svc s
C1Fsvc1 (s)
bsvc
K4
Kp
Kq
Kv
- -
1
K 3 + sTd0 '
K5
KA
1 + sTA
K6
Fig. 3.17 Linearized model with SVC voltage control function included
From Eq. (3.67) and Fig. 3.3, linearized HeffronPhillips model of the power
system with the SVC installed is shown in Fig. 3.17, where the SVC voltage
controller adopts a PI controller and usvcs is the stabilizing signal of the SVC-based
stabilizer.
:
0
0 Kvi
In Eq. (3.66), denote Dzsvc Cxa
s C1 Dd C2 DEq C3 Dbsvc , that is
svcl
:
Ca0
D z
Kvi C1 Dd C2 DE0q C3 Dbsvc
svc
xsvcl
:
3:68
Equation (3.66) becomes

:
Dbsvc

_ 0
Ca
Ca0

Kvp C1 Dd C2 DE0q C3 Dbsvc Dz
Dusvcs 3:69
xsvcl
xsvcl
110
Dbsvc

_
C_ a0
0
0
Cxa
K
C
Dd
C
DE
vp
1
2
q Dz xsvcl Dusvcs
svcl
_
0
1 Cxa
Kvp C3
svcl
3:70
Substituting Eq. (3.70) into Eq. (3.68), it can have

D_zsvc CA Dd CE DE0q CZ Dz CU Dusvcs
3:71
where
0
1
:
Ca0 K C
Ca0
B
xsvcl vp 3 C
CA
Kvi C1 @1
:
A;
xsvcl
0
1 Cxa
K
C
vp 3
svcl
0
1
:
:
Ca0 K C
Ca0
B
xsvcl vp 3 C
CE
Kvi C2 @1
:
A;
xsvcl
0
1 Cxa
K
C
vp 3
svcl
:
Ca0 K C
xsvcl vi 3
CZ
;
:
0
1 Cxa
K
C
vp 3
svcl
: 2
Ca0 K C
vi 3
xsvcl
CU
:
0
1 Cxa
Kvp C3
svcl

:
D d xo Dx
1
:
D x K01 Dd DDx K02 DE0q Kpz Dz K0p Dusvcs
M
1
0
DE_ q 0 K03 DE0q K04 Dd Kqz Dz K0q Dusvcs DEfd
Td0
1
KA 0
0
0
DE_ fd DE_ fd
K Dd K06 DE0q Kvz Dz K0v Dusvcs
TA
TA 5
where
:
0
Cxa
Kvp C1
0
svcl
K1 K1 Kp
;
:
0
1 Cxa
K
C
vp
3
svcl
Kp
Kpz
Ca0 K C
xsvcl vp 3
K0p
0
Cxa
Kvp C2
0
svcl
K2 K2 Kp
;
:
0
1 Cxa
K
C
vp
3
svcl
:
C
a0
Kp xsvcl
:
Ca0 K C
xsvcl vp 3
3:72

:
111
:
0
Cxa
Kvp C2
0
svcl
K3 K3 Kq
;
:
0
1 Cxa
K
C
vp
3
svcl
Kq
Kqz
K0q
0
Cxa
Kvp C1
0
svcl
K4 K4 Kq
;
:
0
1 Cxa
K
C
vp
3
svcl
:
C
a0
Kq xsvcl
Ca0 K C
xsvcl vp 3
Ca0 K C
xsvcl vp 3
K05 K5 Kv
Kvz
0
Cxa
Kvp C1
svcl
:
0
1 Cxa
Kvp C3
svcl
Kv
:
Ca
1 x 0 Kvp C3
svcl
K06 K6 Kv
0
1 Cxa
Kvp C3
svcl
0
Kv Cxa
svcl
K0v
0
Cxa
Kvp C2
svcl
Ca
1 x 0 Kvp C3
svcl
By writing Eqs. (3.71) and (3.72) together in matrix form, the state equation of
the system can be obtained to be
X_ AX bDusvcs
3:73
where
3
0
0
6
K 7
7
6 p 7
6
M 7
6
6 Dx 7
7
6
6
K0q 7
7
6
6 DE0 7
q 7; b 6 0
X6
7;
T
7
6
6
do 7
7
6
6 DE0fd 7
5
6 KA K0v 7
4
4 T 5
A
Dzsvc
CU
2
0
xo
0
0
0
0
6
K
K
D
1
2
6
M
M
0
6
M
6
0
0
K
K
6
1
0
03
A 6 T0 4
Tdo
T0do
6
do
6
K K0
6 KA K05
0
TA 6 T1
4
TA
A
A
CA
0
CE
0
2
Dd
3
0
K 7
Mpz 7
7
7
Kqz 7
0 7
Tdo 7
7
K
A Kvz 7
T
5
A
CZ
112
Linearization Eq. (3.63) is

Dusvcs Ks
1 sT2 1 sT4
DPt Tsvcs sDPt
1 sT1 1 sT3
3:74
Tsvcs sK01 Dd K02 DE0q Kpz Dz K0p Dusvcs

Hence, the output equation of the system is as follows:
y cX dDusvcs
3:75
where

y DPt ; c K01
3.3.1.2
K02

Kpz ; d K0p
Linearized Model of the Example Power System
At the given steady-state operating condition, the voltage at the busbar where the
SVC is installed is Vs0 1:0 p:u: This is achieved by an initial compensation from
the SVC at the operating condition, bsvc0 . This initial compensation can be calculated as follows as introduced in Sect. 3.1.1.4.
From Eqs. (3.24) and (3.26), the reactive power injection at the busbar with the
SVC installed is as follows
Qsb0
Qts0
2
3
s

2
Vs0 4
x
sb
Vs0 V2b
Psb0 5 0:0377
xsb
Vs0
2s
3

2
Vs0 4
x
ts
V2t0
Pts0 Vs0 5 0:0377

xts
Vs0
Hence, the initial compensation is calculated to be

bsvc0
Qsb0 Qts0
0:0754
V2s0
This gives csvc0 1 xsb bsvc0 0:9774.

bsvc0 1 C(a0 =Xsvcl 1=Xsvcc
113
Thus, it can be obtained that at the steady-state operating condition,

C(a0
2a0 sin 2a0

1 0:0754
2p
Hence, a0 42:0727 .
From the circuit model of the power system installed with the SVC in Fig. 3.5, it
can be calculated that
Its0 Pt0 jQts0 0:5 j0:0377
Vs0

Its0 Is0 0:5 j0:0377; Vb0 Vs0 jIsb0 xsb 0:9887 j0:15
Is0 Vs0 jbSVC0 j0:0754;

Isb0
Psb0 jQb0 Isb0 Vb0 0:4887 j0:1123

With the reference voltage, recalculation of the value of various variables is as
follows
Isb0 Pt0 jQb0 0:5 j0:0377; V
s0 V
b0 jXsbIsb0 0:9887 j0:15
b0
V
Is0 V
s0 jbsvc0 0:0113 j0:0746; Its0 Isb0 Is0 0:4887 j0:1123
t0 V
s0 jIts0 xts 0:955 j0:2966; E
Q0 V
t0 jxqIts0 0:8652 j0:6876
V
d0 38:4741

Id0 sin d0 j cos d0 Its0 0:2161; Eq0 EQ0 xd xq Id0 1:1483

E0q0 EQ0 x0d xq Id0 0:9646; Isbd0 sin d0 j cos d0 Isb0 0:2816
Isbq0 cos d0 j sin d0 Isb0 0:4149
From Eq. (3.7) it can have
xdR
xts
0:6069; x0dR x0d xlR 0:7569;

cSVC0
xd xlR 1:6069; xqR xq xlR 1:4069
xlR xts
Thus, from Eqs. (3.9) and (3.12), parameters of the HeffronPhillips model of
the example power system can be calculated to be
K1 0:8765;
K2 0:8410;
K3 2:1229;
K6 0:7475;
Kp 0:0166;
Kq 0:2990;
K4 0:7148;
Kv 0:0806
K5 0:0473;
114

CA 0:0358;
K01 0:8766;
CZ 0:4259; CE 0:7615; CU 0:1218;

K02 0:8394; Kpz 0:0175; K0p 0:0050; K03 2:1514;
K04 0:7134;
Kqz 0:3158;
K06 0:7398;
Kvz 0:0852;
K0q 0:0903;
K05 0:0476;
K0v 0:0243:
State matrix of Eq. (3.73) thus is obtained to be

2
0
6 0:1461
6
A6
6 0:1415
4 95:2539
0:0358
314:2
0
0
0
0
0
0:1399
0:4299
1479:6
0:7614
0
0
0:1983
100:0
0
3
0
0:0029 7
7
0:0626 7
7
170:3097 5
0:4259
Eigenvalues of the state matrix are calculated to be

2;3 0:1349 j6:7166
k1 96:9595; k
4 3:2758; k
5 0:3474
k
Hence, the electromechanical oscillation mode of the power system is
2;3 0:1349 j6:7166.
k
3.3.2
Design of SVC-Based Stabilizer
3.3.2.1
SVC Stabilizer Designed by Using the Phase Compensation

Method
From Fig. 3.17 and Eq. (3.74), the forward path of the stabilizing signal of the
SVC-based stabilizer, D usvc-s, can be drawn as shown in Fig. 3.18. Transfer
function of the forward path of the SVC-based stabilizer can be obtained to be

9
8

=
<
Kq Kv 1 +KAsT C2 Fsvc1 s)Kp K2
A

Fsvcs s) Fsvc2 s) Kp
;
:
A
A
K3 sT0d0 K6 1 KsT
C2 Fsvc1 s) Kq Kv 1 KsT
A
A
3:76
Hence, from Fig. 3.17, the electric torque contribution from the SVC-based
stabilizer can be obtained to be
DTet Fsvcs s)Tsvcs s)( DPt Fsvcs s)Tsvcs s)(D sM)Dx
3:77
115
K1
Pt
1
D + sM
0
s
Tsvc s (s)
K2
u svc s
C2 Fsvc1 (s)
Fsvc2 (s)
bsvc
Kp
1
K 3 + sTd0 '
Kq
Kv
KA
1 + sTA
K6
Fig. 3.18 Forward path of the SVC-based stabilizer
2;3 0:1349 j6:7166. By taking

Oscillation modes of the power system are k
xs 6:7166, from Eqs. (3.67) and (3.76), it can have
svc1 jxs 0:2899 j0:3040;
F

Fsvcs jxs 0:0912 j0:0114
svc2 jxs 0:2696 j0:0171;

F
Take the transfer function of the SVC stabilizer to be

Tsvcs s) Ks
1 sT2 1 sT4
;
1 sT1 1 sT3
T1 0:09 s:; T3 0:09 s:
116
The SVC stabilizer is to be designed to provide a damping torque

Dsvc Dx 1:6Dx, i.e., Dsvc 1:6. Parameters of the SVC-based stabilizer can be
set by using the phase compensation method to provide a pure damping torque
DTet Fsvcs jxs D jxs M)Tsvcs jxs Dx Dsvc Dx
3:78
The results of computation by the use of Eq. (3.78) are as follows

Ks K1 K2 5:8540;
T2 1:7369 s;
T4 1:7369 s
Let
Dx1
1 sT4
Dx;
1 sT3
Dusvcs K
1 sT2
Dx1
1 sT1
Thus,
h

T4 K01
4
sDx1 T1 Dx1 T1 M
Dd DT

1
Dx
M
3
3

0
0
T4 K
T4 K2 0 T4 Kpz
M
DEq M Dzsvc M p Dusvcs
sT2 Dx
sDusvc s Ks11sT
1
1
3:79
From Eqs. (3.73) and (3.79), state equation of the closed-loop system can be
obtained to be
2
6
6
6
6
6
6
6
6
6
4
Dd
:
Dx
0
DE_ q
0
DE_ fd
D_zsvc
Dx_ 1
7
6
7
6
7
6
7
6
7
7 Ac 6
6
7
6
7
6
7
4
5
Du_ svcs
Dd
Dx
DE0q
DE0fd
Dzsvc
Dx1
Dusvcs
3
7
7
7
7
7
7
7
7
5
3:80
where
2
0
314:1593
6 0:1461
0
6
6 0:1415
0
6
Ac 6
0
6 95:2539
6 0:0358
0
6
4 0:7364
524:6004
7:4877
5333:9
0
0
0:1399
0
0:4265 0:1983
1479:6 100
0:7615
0
0:2133 0:3170
2:1683 3:2233
0
0:0029
0:0626
170:3097
0:4259
0:0005
0:0056
3
0
0
0
0:0008 7
7
0
0:0179 7
7
0
48:6868 7
7
0
0:1218 7
7
1:0411
0:0205 5
4:7312 0:7913

Fig. 3.19 Simulation of the
example power system
without and with the SVC
stabilizer installed
42
/degree
117
Without SVC stabilizer
With SVC stabilizer
41
40
39
38
37
36
35
t/s
0
10
Eigenvalues of the state matrix of the closed-loop system are as follows

k1 95:2487;
k5 1:2281;
2;3 0:6870 j6:6669;

k
k6 0:2937;
k4 3:7484;
k7 0:7921
2;3
The electromechanical oscillation mode of the power system is now k
0:6870 j6:6669 with sufcient damping.
Figure 3.19 gives the simulation results of the power system without and with
the SVC stabilizer installed. At 0.5 s of the simulation, a three-phase to-earth short
circuit occurred on the transmission line of the single-machine innite-bus power
system in Fig. 3.2. The fault was cleared in 0.1 s. From Fig. 3.19, it can be seen that
the SVC stabilizer effectively suppresses the power oscillation.
3.3.2.2
Damping Control Effectiveness of the SVC Stabilizer

as Affected by Various Factors
Effectiveness of the SVC stabilizer installed is affected by various factors as

discussed in Sect. 3.1.2. Figure 3.20 shows the computational results of the
damping torque provided by the SVC stabilizer when the load condition of the
power system varies. From Fig. 3.20, it can be seen that
1. The direct damping torque provided by the SVC stabilizer is much greater than
the indirect damping torque;
2. When the active power supplied by the generator increases, more damping
torque is provided by the SVC stabilizer.
Table 3.1 shows the computational result of electromechanical oscillation mode
of the power system with the SVC stabilizer installed at three different load
118
Fig. 3.20 Computational

results of the damping torque
provided by the SVC
Damping torque provided

by the SVC stabilizer (p.u.)
45
40
Total damping torque
35
30
25
20
15
Direct
damping torque
10
Indirect damping torque
5
0
Pt0 (p.u.)
-5
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Table 3.1 Electromechanical oscillation mode

Load condition Pt0 (p.u.)
0.3
0.5
0.7
Electromechanical
oscillation mode
0:1732 j5:4744
0:6870 j6:6669
1:0831 j6:0391
conditions. Figures 3.21 and 3.22 show the simulation results. All those results in
Figs. 3.20, 3.21, and 3.22 and in Table 3.1 conrm the analytical conclusion that
the effectiveness of the SVC stabilizer increases with the load condition of power
system.
Fig. 3.21 Simulation results
at Pt0 0:3
/degree
Without the SVC

stabilizer
With the SVC stabilizer
27
26
25
24
23
22
21
20
t/s
0
10

at Pt0 0:7
58
119
Without the SVC stabilizer
/degree
With the SVC stabilizer
56
54
52
50
48
t/s
46
0
10
Discussions in Sect. 3.1.2 conclude that the effectiveness of the SVC stabilizer is
also affected by the length of the transmission line and the installing locations of the
SVC stabilizer. To conrm the conclusion, Fig. 3.23 presents the computational
results of the damping torque provided by the SVC stabilizer at variable load
condition when the system operates with single and double transmission lines
connecting the generator and the innite busbar. From Fig. 3.23, it can be seen that
when the electric length of the line is longer (single transmission line connecting the
180
Damping torque provided by

the SVC stabilizer(p.u.)
Single line
160
140
120
100
Double
lines
80
60
40
20
Pt0 (p.u.)
0
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Fig. 3.23 Computational results of the damping torque provided by the SVC stabilizer
120
12
Direct damping
torque
Damping torque provided

by the SVC stabilizer
10
Total damping
torque
Indirect damping
torque
0
Xts+Xsb
-2
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Fig. 3.24 Computational results of the damping torque provided by the SVC stabilizer installed at
various locations
generator and the innite busbar), the SVC stabilizer provides more damping torque
and hence is more effective in suppressing the power oscillation.
Keeping the operating condition of the example power system unchanged at
Pt0 0:5; Vb0 1:0; Vt0 1:0; Vs0 1:0, the computational results of the
damping torque provided by the SVC stabilizer installed at various locations along
the transmission line are presented in Fig. 3.24. From Fig. 3.24, it can be seen that
the SVC stabilizer is most effective when it is installed around the middle point of
the line as it provides the most damping torque at the point.
References
1. Wang HF, Swift FJ (1996) Capability of the static VAr compensator in damping power system
oscillations. IEE Proc Part C 143(4):353358
2. Smith OJM (1969) Power system transient control by capacitor switching. IEEE Trans Power
Apparatus Syst 88(1):2835
3. Smith OJM, Webster RH (1971) Series capacitor switching to quench electromechanical
transients in power systems. IEEE Trans Power Apparatus Syst 90(2):427433
4. Wang HF, Swift FJ (1996) Application of the Controllable Series Compensator in damping
power system oscillations. IEE Proc Part C 143(4):359364
5. Wang HF, Swift FJ (1997) Analysis of thyristor-controlled phase shifter applied in damping
power system oscillations. Int J Electr Power Energ Syst 19(1):19
Chapter 4
Single-Machine Innite-Bus Power

Systems Installed with VSC-Based
Stabilizers
4.1
Damping Torque Analysis of a Shunt VSC-Based

Stabilizer Installed in a Single-Machine Innite-Bus
Power System
4.1.1

Innite-Bus Power System Installed with a Shunt
VSC-Based Stabilizer
4.1.1.1
A Shunt VSC Installed in a Single-Machine Innite-Bus Power

System
Figure 4.1 shows the conguration of a shunt voltage source converter (VSC)-based
unit connected to high-voltage transmission line through a step-down transformer.
xs is the equivalent reactance of the step-down transformer. The unit can be an
energy storage system (ESS), such as a battery energy storage system (BESS) or a
superconducting magnetic energy storage (SMES) device. It can also be a renewable power generation plant, such as a grid-connected photovoltaic (PV) or fuel cell
(FC) power plant. In case that Idc2 0, it is a static synchronous compensator
(STATCOM) [1].
There are two commonly used algorithms to control the voltage at the AC
terminal of the converter Vc , the pulse width modulation (PWM), and the pulse
amplitude modulation (PAM). When the PWM is used, Vdc is kept constant by
controlling the modulation phase, (angle between Vc and Vs ), to charge or discharge the capacitor. The amplitude of voltage at the AC terminal is regulated by

DOI 10.1007/978-1-4899-7696-3_4
121
4 Single-Machine Innite-Bus Power Systems
122
Fig. 4.1 Conguration of a
shunt VSC-based unit
transmission line
Vs
Is
xs
Vc
Idc
Cdc
Idc1
Idc2
Vdc
the modulation ratio m, i.e. Vc mkVdc , where k is the ratio between the AC and
DC voltages of the converter (including the ratio of the step-down transformer),
dependent of the structure of the converter circuit.
Figure 4.2 shows the conguration of a single-machine innite-bus power
system, where a shunt VSC-based unit is installed at a location along the transmission line. From Fig. 4.2, it can be obtained that
Vt jxts Its Vs
4:1
Vs jxs Is Vc
Vs Vb jxsb Its Is
Fig. 4.2 A single-machine

installed with a VSC-based
unit
Vt
x ts
Vs
x sb
Isb
Its
Is
xs
Vc
VSC-based unit
Vb
123
Hence
jxs Is Vc Vb jxsb Its Is
4:2
Vt jxts Its jxsb Its Is Vb

In the dq coordinate of generator, as shown in Fig. 4.3, Eq. (4.2) gives
jxs isd jisq Vc cos w j sin w Vb sin d j cos d
jxsb [(itsd isd j(itsq isq
4:3
vtd jvtq
jxts xsb itsd jitsq jxsb isd jisq Vb sin d j cos d
By equating the real and imaginary parts on both sides of Eq. (4.3), it can have
xs isq Vc cos w Vb sin d xsb itsq isq
vtd xts xsb itsq xsb isq Vb sin d
xs isd Vc sin w Vb cos d xsb itsd isd
4:4
vtq xts xsb itsd xsb isd Vb cos d

Because
vtd xq itsq
vtq E0q x0d itsd
Fig. 4.3 Phasor diagram of

system shown in Fig. 4.2
4:5
Vt
Vs
jxs Is
jxts Its
Vc
Vb
jxsb Isb
124

xsb
xs xsb
xq xts xsb
xsb
xsb
xs xsb
x0d xts xsb
xsb

itsq
isq
Itsd
Isd
"
Vc cos w Vb sin d

Vb sin d
Vc sin w Vb cos d
E0q Vb cos d
4:6

itsq
isq
xsb
xq xts xsb

xsb
itsd
0
xd xts xsb
isd
xs xsb
xsb
1
Vc cos w Vb sin d

Vb sin d
#
1 "
Vc sin w Vb cos d
xs xsb
xsb

4:7
E0q Vb cos d
The active power supplied by the generator can be expressed as

Pt vtd itsd vtq itsq xq itsq itsd E0q x0q itsd itsq
E0q itsq xq x0q itsd itsq
4:8

Eq E0q xd x0q itsd
q q
Vt v2td v2tq xq itsq 2 E0q x0q itsd 2
4:9
Magnitude of Vc is mkVdc . In the dq coordinate of generator as shown in Fig. 4.3,

its phase is w. Hence,
Vc mkVdc cos w j sin w mkVdc \w
4:10
Active power received by the DC capacitor of VSC-based unit is

Vdc Idc1 isd vcd isq vcq isd mkVdc cos w isq mkVdc sin w
4:11

Idc isd mk cos w isq mk sin w
4:12
Dynamic equation on the DC side of the converter is

1
1
Idc2
V_ dc
Idc1 Idc2
isd mk cos w isq mk sin w
Cdc
Cdc
Cdc
4:13
125
From the phasor diagram of Fig. 4.3, it can have

w tan1
vsq
/
vsq
4:14
Equations (4.7)(4.9), (4.13), and (4.14), together with the differential equations of
synchronous generator of Eq. (2.37), form the mathematical model of the
single-machine innite-bus power system with the shunt VSC-based unit installed.
In Eqs. (4.10) and (4.14), the modulation ratio and phase of PWM algorithm, m and
/, are regulated by two controllers. Hence, m and / are two control variables in the
mathematical model of the power system.
4.1.1.2
Extended HeffronPhillips Model
Linearization of Eq. (4.7) is

Ditsq
Disq
c11
c12
c21 c22

m0 k cos w0 DVdc kVdc0 cos w0 Dm m0 kVdc0 sin w0 Dw Vb cos d0 Dd

Vb cos d0 Dd

d11 d12
Ditsd
Disd
d21 d22
"
#
m0 k sin w0 DVdc kVdc0 sin w0 Dm m0 kVdc0 cos w0 Dw Vb sin d0 Dd

DE0q Vb sin d0 Dd
4:15
The second equation in Eq. (4.1) is
vsd jvsq jxs isd jisq mkVdc cos w jmkVdc sin w
Hence, it can have
vsd xs isq mkVdc cos w
vsq xs isd mkVdc sin w
By using Eq. (4.15), linearization of above equations can be obtained as
Dvsd xs 1m0 k cos w0 DVdc xs 1kVdc0 cos w0 Dm
xs 1m0 kVdc0 sin w0 Dw xs Vb cos d0 Dd
ad1 DVdc ad2 Dm ad3 Dw ad4 Dd
Dvsq xs 1m0 k sin w0 DVdc xs 1kVdc0 sin w0 Dm
xs 1m0 kVdc0 cos w0 Dw Vb sin d0 D
aq1 DVdc aq2 Dm aq3 Dw aq4 Dd
126
By using above equations, linearization of Eq. (4.14) can be obtained as

Dw a01 DVdc a02 Dm a03 Dw a04 Dd D/
where
1
vsd0 a1q vsq0 a1d
v2sd0 v2sq0
1
a02 2
vsd0 a2q vsq0 a2d
vsd0 v2sq0
1
a03 2
vsd0 a3q vsq0 a3d
vsd0 v2sq0
1
a04 2
vsd0 a4q vsq0 a4d
vsd0 v2sq0
a01
That is
Dw
a01
a02
a04
1
Dd
D/
0 DVdc
0 Dm
1 a03
1 a3
1 a3
1 a03
a1 DVdc a2 Dm a3 Dd a4 D/
4:16
By using Eqs. (4.15) and (4.16), linearization of Eqs. (4.8) and (4.9) can be
obtained as
DPt K01 Dd K2 DE0q K0pdc DVdc K0pm Dm K0pw Dw
K01 a3 K0pw Dd K2 DE0q K0pdc a1 K0pw DVdc
K0pm a2 K0pw Dm a4 K0pw D/
K1 Dd K2 DE0q Kpdc DVdc Kpm Dm Kpw D/
DEq K04 Dd K3 DE0q K0qdc DVdc K0qm Dm K0qw Dw
K04 a3 K0qw Dd K3 DE0q K0qdc a1 K0qw DVdc
K0qm a2 K0qw Dm a4 K0qw D/
K4 Dd K3 DE0q Kqdc DVdc Kqm Dm Kqw D/
DVt K05 Dd K6 DE0q K0vdc DVdc K0vm Dm K0vw Dw
K5 a3 K0vw Dd K6 DE 0 q K0vdc a1 K0vw DVdc
K0vm a2 K0vw Dm a4 K0vw D/
K5 Dd K6 DE0q Kvdc DVdc Kvm Dm Kvw D/
4:17
127
where
C0 m0 k
h

i
K01 E0q0 itsd0 x0d xq c11 c12 Vb cos d0

itsq0 x0d xq d11 d12 Vb sin d0

K2 itsq0 itsq0 x0d xq d12
h

i

K0pdc c11 C0 cos w0 E0q0 itsd0 x0d xq d11 C0 sin w0 Itsq0 x0d xq
h

i

K0pm c11 kVdc0 cos w0 E0q0 itsd0 x0d xq d11 kVdc0 sin w0 Itsq0 x0d xq
h

i

K0pw c11 C0 Vdc0 sin w0 E0q0 itsd0 x0d xq d11 C0 Vdc0 cos w0 Itsq0 x0d xq
K4 d11 d12 Vb sin d0 xd x0d
K03 d12 xd x0d 1
K0qdc d11 C0 sin w0 xd x0d
K0qm d11 kVdc0 sin w0 xd x0d
K0qw d11 C0 Vdc0 cos w0 xd x0d

vtd0 Xq c11 c12 Vb cos d0 vtq0 x0d d11 d12 Vb sin d0
0
K5
Vt0

vtq0 1 x0d d12
K6
Vt0

v
xq c11 C0 cos w0 vtq0 x0d d11 C0 sin w0
td0
0
Kvdc
Vt0

v
x
kc
V
cos
w0 vtq0 x0d d11 kVdc0 sin w0
td0
q
11
dc0
0
Kvm
Vt0

vtd0 xq c11 C0 Vdc0 sin w0 vtq0 x0d d11 C0 Vdc0 cos w0
0
Kvw
Vt0

Cdc DV_ dc isd0 k cos w0 isq0 k sin w0 Dm
m0 kisd0 sin w0 isq0 cos w0 Dw
m0 kcos w0 Disd sin w0 Disq DIdc2
128
By using Eqs. (4.15) and (4.16), it can have

DV_ dc K07 Dd K8 DE0q K09 DVdc K0dm Dm + K0dw Dw
1
DIdc2
Cdc
K07 a3 K0dw Dd K8 DE0q K09 a1 K0dw DVdc

1
Kdm a2 K0dw Dm + a4 K0dw Df +
DIdc2
Cdc
K7 Dd K8 DE0q K9 DVdc Kdm Dm + Kdw D/
1
DIdc2
Cdc
4:18
where
m0 k cos w0 d21 d22 Vb sin d0 m0 k sin w0 c21 c22 Vb cos d0
Cdc
m0 k cos w0 d22
K8
C
2 2 dc

m
k
cos w0 d21 sin w0 m20 k2 sin w0 c21 cos w0
0
0
K9
Cdc
1
K0dm
kisd0 cos w0 kisq0 sin w0 m0 k2 cos w0 d21 Vdc0 sin w0
Cdc
m0 k2 sin w0 c21 Vdc0 cos w0
1
K0dw
m0 kisd0 sin w0 isq0 cos w0 m20 k2 cos w0 d21 Vdc0 cos w0
Cdc
m20 k2 sin w0 c21 Vdc0 sin w0
K07
By substituting Eq. (4.17) into the linearized differential equation of synchronous generator of Eq. (2.37) (for the simplicity of expression, the PSS is not
considered), the extended HeffronPhillips model of the single-machine innite-bus
power system installed with the shunt VSC-based unit can be obtained as
Dd_ xo Dx
1
Dx_ K1 Dd DDx K2 DE0 q Kpdc DVdc Kpm Dm Kpw D/
M
1
DE_ 0q 0 K4 Dd K3 DE0 q DE0fd Kqdc DVdc Kqm Dm Kqw D/
Td0
1
KA
K5 Dd K6 DE0 q Kvdc DVdc Kvm Dm Kvw D/
DE_ 0fd DE0fd
TA
TA
1
DV_ dc K7 Dd K8 DE0 q K9 DVdc Kdm Dm Kdw D/
DIdc2
Cdc
4:19
129
The model is shown in Figs. 4.4 and 4.5.
Idc2
E q'
K8
K7
1
Cdc
1
s K9
K dm
K d
Vdc
Fig. 4.4 Extended HeffronPhillips modelpart of the VSC-based unit
K 1 K P
1
sM + D
0
s
K 4 K q K 5 K v
K2
Vdc
[K pdc K p K pm ]
[ ]
m
[K vdc K v K vm ]
[K qdc K q K qm ]
E q '
1
Td0 's + K 3
KA
1 + sTA
K6
Fig. 4.5 Extended HeffronPhillips modelpart of the power system
130
4.1.2
Damping Torque Analysis of Shunt VSC-Based

Stabilizer Installed in Single-Machine Innite-Bus
Power System
4.1.2.1
Damping Torque Contributed by the Shunt VSC-Based

Stabilizer [2, 3]
A damping control function can be superimposed on the modulation ratio m of the

PWM algorithm. This forms a VSC-based stabilizer which functions through regulating the exchange of reactive power between the VSC-based unit and rest of the
power system. In this book, it is named as the VSC-based reactive power stabilizer
for the convenience of discussion. If it is assumed that the feedback signal and
transfer function of the stabilizer are y and Tvsc s, respectively, and only the
stabilizers function is considered, it can have Dm Tvsc sDy. At the angular
oscillation frequency of the power system, xs , stabilizing signal of the VSC-based
reactive power stabilizer Dm Tvsc jxs Dy can always be decomposed into a
synchronizing and damping torque, that is
Dm Tvsc jxs Dy Kvscd Dx Kvscs Dd
4:20
For the simplicity of discussion and presentation, it is assumed that the feedback
signal of the stabilizer is the rotor speed of generator and the stabilizer adopts a
proportional control law, it has
Dm Krs Dx
4:21
From Eq. (4.20), it can be seen that the above assumption will not change the
following discussion and conclusions about the effectiveness of stabilizer.
From Eq. (4.21), Figs. 4.4, and 4.5, the direct damping torque contribution from
the VSC-based reactive power stabilizer can be obtained as
DTrsd

Kpm Kpdc Kdm
K9
Krs Dx
x2s K29
4:22
If the shunt VSC-based unit (such as an energy storage system or a renewable

power generation plant) can exchange active power with the rest of power system, a
damping control function can also be superimposed on the modulation phase of
PWM algorithm, /. This VSC-based stabilizer functions by controlling the
exchange of active power between the VSC-based unit and rest of the power
system. In this book, it is named as the VSC-based active power stabilizer. Again, it
can be assumed that the feedback signal of stabilizer is the rotor speed of generator
and the stabilizer is a pure-gain controller. That is
D/ Kas Dx
131
4:23
From Eq. (4.23), Figs. 4.4, and 4.5, the direct damping torque contribution from the
VSC-based active power stabilizer can be obtained as

DTasd
K9
Kpw Kpdc Kdw 2
Kas Dx
xs K29
4:24
In the following section, difference between the damping effectiveness of the

VSC-based reactive and active power stabilizers is examined using Eqs. (4.22) and
(4.24).
4.1.2.2
Difference of Damping Control Effectiveness Between

the VSC-Based Reactive and Active Power Stabilizers

VS
jxsb
xsb
Vb
Its
Vc
x
x
sb
sb
1 xs
xs 1 xs
1 xxsbs
4:25
That gives
Vt jxts Its Vs

xs xsb
xsb
xs
j xts
Its
Vc
Vb jxIts Va
xs xsb
xs xsb
xs xsb
4:26
where

x
Va
xts
xs xsb
xs xsb
xsb
xs
Vc
Vb aVc bVb
xs xsb
xs xsb
Hence, from Eq. (4.26), the active power supplied by the generator can be
expressed as
Pt
E0q Va
V2a xq x0d
0
sin
d

sin 2d0
x0dR
2 x0dR xqR
4:27
where d0 is the angle between E0q and Va and x0dR x x0d ; xqR x xq . From the
phasor diagram of Fig. 4.6, it can have
132
Fig. 4.6 Phasor diagram for
Eq. (4.27)
q
E
Va
aVc
bVb
d
Va sin d0 bVb sin d aVc cos w
Va cos d0 bVb cos d aVc sin w
4:28
Thus, from Eqs. (4.27) and (4.28), the active power supplied by the generator can
be expressed alternatively to be
Pt
E0q
bVb sin d aVc cos w

x0dR
xq x0d
0
bVb sin d aVc cos wbVb cos d aVc sin w
xdR xqR
E0q
0 bVb sin d aVc cos w
xdR
4:29
where Vc mkVdc . Equation (4.29) is the explicit expression of the active power
supplied by the generator in the single-machine innite-bus power system installed
with the shunt VSC-based unit. By using Eqs. (4.16) and (4.17), from Eq. (4.29), it
can be obtained that

@Pt
a4 akVdc0 cos w0
@m0

@Pt
Kpw a4 a4 akm0 Vdc0 sin w0
@w 0

@Pt
Kpdc a4
a4 akm0 cos w0
@Vdc 0
Kpm a4
4:30
133
From Eqs. (4.22) and (4.30), the damping torque contributed by the VSC-based
reactive power stabilizer is obtained as

DTrsd a4 ak cos w0 Vdc0 m0 Kdm
K9
Krs Dx
x2s K29
4:31
From Eqs. (4.24) and (4.30), the damping torque contributed by the VSC-based
active power stabilizer is derived as

DTasd a4 akm0 Vdc0 sin w0 cos w0 Kdw
K9
Kas Dx
x2s K29
4:32
From Fig. 4.3, it can be seen that d0 increases when more active power is supplied
by the generator. Hence, at a heavier load condition, w0 is smaller, and more
damping torque is provided by the VSC-based reactive power stabilizer as indicated
by Eq. (4.31). This means that the VSC-based reactive power stabilizer is more
effective in damping power oscillation when the single-machine innite-bus power
system operates at the heavier load condition. However, with the increase of load
condition, cos w0 increases, but sin w0 decreases in Eq. (4.32). This means that with
variations of power system load conditions, the damping torque provided by the
VSC-based active power stabilizer changes less than that provided by the reactive
power stabilizer. Hence, the VSC-based active power stabilizer is more robust to the
variations of power system load conditions in damping power oscillations.
4.2
Damping Function of a Stabilizer Added on a Static

Synchronous Series Compensator (SSSC) Installed
in a Single-Machine Innite-Bus Power System
4.2.1
Damping Torque Analysis of a SSSC Stabilizer

Installed in a Single-Machine Innite-Bus Power
System
4.2.1.1

with a Static Synchronous Series Compensator
A static synchronous series compensator (SSSC) is a series STATCOM, a

solid-state voltage source inverter generating a controllable AC voltage source, and
connected in series to power transmission lines in a power system. The injected
voltage is in quadrature with the line current and emulates an inductive or capacitive reactance so as to influence the power flow along the transmission lines. With
the capability to change its reactance characteristic from capacitive to inductive, the
SSSC is very effective in controlling power flow in the power system. An auxiliary
134
stabilizing signal can be superimposed on the power flow control function of the
SSSC so as to improve the damping of power system oscillations.
Figure 4.7 shows a single-machine innite-bus power system installed with a
SSSC, which consists of a series coupling transformer with a leakage reactance,
xsct , and a VSC connected to a DC capacitor. The exchange of reactive power
between the SSSC and rest of the power system is achieved by controlling the
magnitude and phase of the inserted voltage which is kept in quadrature with the
line current (the inverter losses are ignored). Phasor diagrams in Fig. 4.8 show the
basic operation principle of the SSSC, from which it can be seen that the compensation level can be controlled dynamically by regulating the magnitude of Vinv .
Mathematical description of the SSSC is
Vt
xt
It
Pt
I ts
x ts
V sssc
V1
Vb
+ V inv -
x sct
V2
x sb
Pts
VSC
Idc
Vdc
+
Fig. 4.7 A single-machine innite-bus power system with a SSSC
V inv
V inv
V2
V1
jx cst I ts
jx cst I ts
V2
I ts
SSSC in inductive mode
Fig. 4.8 Phasor diagram of operation of the SSSC
V1
I ts
SSSC in capacitive mode
4.2 Damping Function of a Stabilizer Added on a Static
135
Its itsd jitsq Its \u

Vinv mkVdc cos / j sin / mkVdc \/;
dVdc
Idc
mk
itsd cos / itsq sin /

dt
Cdc Cdc
u / 90
4:33
where k is the xed ratio between the inverters AC and DC voltages and m is the
modulation ratio of the PWM algorithm implemented by the series VSC.
According to Fig. 4.8, a variable of equivalent reactance, xdc , can be introduced
to represent the AC voltage, Vinv , to be
Vinv jxdc Its
4:34
From the circuit equation of power system of Fig. 4.7, it can have
Vt Vb j(xts xsb xsct Its Vinv jxt It
4:35
In the dq coordinate, the above equation becomes

xts xsb xsct itsq j(xts xsb xsct itsd xdc itsq jxdc itsd
xt itq jxt itd
4:36
Hence,
xts xsb xsct xdc
itsq
xt
xts xsb xsct xdc
itd
itsd
xt
itq
4:37
Considering the voltage equation of generator of Eq. (4.35), it can have

Vt j(xts xsb xsct Its xq itsq itq jE0q x0d itsd itd jxts xsb xsct Its

xts xsb xsct xdc
xts xsb xsct xdc
itsq j E0q x0d 1
itsd
xq 1
xt
xt
xts xsb xsct itsq j(xts xsb xsct itsd
Vinv Vb xdc itsq jxdc itsd Vb sin d jVb cos d
4:38
From the above equation, it can be obtained that

Vb sin d
xts xsb xsct cs xq xdc
E0q Vb cos d
xts xsb xsct cs x0d xdc
itsq
itsd
4:39
136
where
cs 1
xts xsb xsct xdc

xt
Thus,
Pt E0q itq itsq xq x0d itd itsd itq itsq
E0q cs itsq xq x0d c2s itsd itsq
Eq E0q xd x0d itd itsd E0q xd x0d cs itsd
q
Vt E0q x0d itd itsd 2 xq itq itsq 2
q
E0q x0d cs itsd 2 xq cs itsq 2
4:40

Vinv jxdc Its xdc itsq jxdc itsd mkVdc cos / j sin /
4:41

xdc itsq
mkVdc
xdc itsd
sin /
mkVdc
cos /
4:42
By using Eq. (4.42), dynamic equation in Eq. (4.33) is converted to the following
equivalent form
dVdc
0
dt
4:43
Equation (4.43) shows that indeed, there is no exchange of active power between
the SSSC and rest of the power system.
4.2.1.2
Extended HeffronPhillips Model of the Single-Machine

Innite-Bus Power System Installed with the SSSC
and Damping Torque Analysis [4]

DPt K1 Dd K2 DE0q Kpx Dxdc
DEq K4 Dd K3 DE0q Kqx Dxdc
DVt
K5 Dd K6 DE0q
Kvx Dxdc
4:44
137

xdc Its mkVdc
4:45
Its0 Dxdc xdc0 DIts kVdc0 Dm
4:46
Hence,
Because only the damping control function will be considered, it can be assumed
that xdc0 0 as far as the damping control function at the steady-state operation is
concerned. Hence, Eq. (4.44) becomes
DPt K1 Dd K2 DE0q Kpm Dm
DEq K4 Dd K3 DE0q Kqm Dm
DVt
K5 Dd K6 DE0q
4:47
Kvm Dm
where
Kpx kVdc0
Its0
Kqx kVdc0
Its0
Kvx kVdc0
Its0
Kpm
Kqm
Kvm
By substituting Eq. (4.47) into Eq. (2.39) (without the PSS being considered), the
extended HeffronPhillips model of the single-machine innite-bus power system
with the SSSC installed can be obtained as
Dd_ xo Dx
1
Dx_ K1 Dd DDx K2 DE0q Kpm Dm
M
1
DE_ 0q 0 K4 Dd K3 DE0q DE0fd Kqm Dm
Td0
1
KA
K5 Dd K6 DE0q Kvm Dm
DE_ 0fd DE0fd
TA
TA
4:48
The model is shown in Fig. 4.9.

Active power supplied by the generator can be expressed as
Pt E0q itsq itq xq x0d itsd itd itsq itq
E0q cs itsq xq x0d c2s itsd itsq
4:49
138
K1
0
s
1
D + sM
K2
K4
K pm
K qm
K vm
- -
1
K 3 + sTd0 '
K5
KA
1 + sTA
K6
Fig. 4.9 Extended HeffronPhillips model of single-machine innite-bus power system installed
with a SSSC stabilizer
By using Eq. (4.39), it can have

Pt
cs E0q Vb
c2s V2b xq x0d
sin
d

sin 2d
x0dR
2 x0dR xqR
4:50
where
x0dR cs x0d xts xsb xsct xdc
xqR cs xq xts xsb xsct xdc
From Eq. (4.50), it can be dened that
cs E0q0 Vb
c2 V2 xq x0d
sin d0 s b 0
sin 2d0 Pt10 Pt20
0
xdR
2 xdR xqR
cs E0q0 Vb
sin d0
x0dR
Pt0
Pt10
Pt20
c2s V2b xq x0d

sin 2d0
2 x0dR xqR
4:51
139
According to Eqs. (4.41) and (4.51), from Eq. (4.50), it can have
Kpx
0

0

xd
xd
xq

xt 1
xt 1
xt 1
@Pt
1
2
Pt10
Pt10
Pt20
Pt20
Pt20
x0dR
xqR
x0
@xdc 0 xt cs
xt c s
0

dR
xd
xq
xL 1
xL 1
1
1
Pt10 Pt20
Pt20
Pt10 Pt20
Pt20
x0dR
xqR
xt cs
xt c s

xq xt
1
x0 xt
1
Pt0
Pt20
d 0

xt xdR
xt xqR
xt c s
xt c s
xqR cs xq xt cs
x0 cs x0d cs xt
dR
Pt0
Pt20
xt cs x0dR
xt cs xqR
1
1
Pt0
Pt20
cs x0dR
cs xqR
4:52
Kpm
kVdc0 1
1

Pt0
Pt20
xqR
cs Its0 x0dR
4:53
Taking the similar procedure of analysis to that from Eqs. (3.58) to (3.59), from Eq.
(4.53), it can be concluded that
1. When the load conditions change, the SSSC stabilizer can always provide
positive damping to power system oscillations if it is designed properly;
2. The heavier the load condition of the power system is, the more effective the
SSSC stabilizer is;
3. When the load condition is unchanged, the SSSC stabilizer is less effective when
the equivalent reactance of transmission line is bigger.
4.2.2
Design of a SSSC Stabilizer
Using a SSSC stabilizer as an example, this section demonstrates how a

thyristor-based or VSC-based stabilizer can be designed effectively to suppress
power system oscillations in a single-machine innite-bus power system using the
phase compensation method.
From Fig. 4.9, the forward path of the SSSC stabilizer is shown in Fig. 4.10.
From Fig. 4.10, the transfer function of the forward path can be obtained as
140
K2
K pm
K qm
K vm
1
K 3 + sTd0 '
KA
1 + sTA
K6
Fig. 4.10 Forward path of stabilizing signal of SSSC stabilizer installed in a single-machine
1
Fsssc s Kpm K2 Kqm
K3 sT0d0
KA
1
K3 sT0d0 1 sTA
KA
1
K3 sT0d0 1 sTA
K2 Kvm
KA
1
1 K6
K3 sT0d0 1 sTA
K2 Kqm 1 sTA Kvm KA
Kpm
K3 sT0d0 1 sTA K6 KA
1 K6
4:54
Consider the case that the feedback signal and transfer function of the SSSC
stabilizer are Dx and Tsssc s), respectively, that is
Dm Tsssc sDx
4:55
If the amount of damping torque contribution required from the SSSC stabilizer is
Dsssc Dx, the transfer function of SSSC stabilizer should be set by using the phase
compensation method at the angular oscillation frequency, xs , to satisfy
Fsssc jxs Tsssc jxs Dsssc
4:56
In a more general case, the state-space representation of linearized model of the

single-machine innite-bus power system with the SSSC stabilizer installed can
always be arranged in the following form:
3 2
0
Dd_
7 6
6
4 Dx_ 5 4 k
A31
x_
2
Dy C1
3 2
3
0
Dd
7 6
76
7
d A23 54 Dx 5 4 B2 5Dm
x
A32 A33
B3
3
2
Dd
7
6
C2 C3 4 Dx 5 dm Dm
x
x0
32
4:57
141
Equation (4.57) can be written as

Dd_ x0 Dx
Dx_ kDd dDx A23 x B2 Dm
x_ A31 Dd A32 Dx A33 x B3 Dm
4:58
Dy C1 Dd C2 Dx CT3 x dm Dm
sDd x0 Dx
sDx kDd dDx A23 x B2 Dm
sx A31 Dd A32 Dx A33 x B3 Dm
4:59
Dy C1 Dd C2 Dx CT3 x dm Dm
Hence,

s
x sI A33 1 A31
A32 Dd sI A33 1 B3 Dm
x0
4:60

sDd x0 Dx

s
s
A23 sI A33 1 A31 A32
sDx k d
Dd
x0
x0
A23 sI A33 1 B3 B2 Dm KsDd Kc sDm

s
Dy C1 Dd C2 Dx CT3 sI A33 1 A31
A32 Dd sI A33 1 B3 Dm
x0
h x

i
0
1 x0
T
C2 C3 sI A33
A31 A32 Dx dm CT3 sI A33 1 B3 Dm
C1
s
s
Ko sDx Kil sDm
4:61
where

s
s
Ks k
d AT23 sI A33 1 A31
A32
x0
x0
x

0
1 x0
T
Ko s
C1 C2 C3 sI A33
A31 A32
s
s
Kc s AT23 sI A33 1 B3 B2
Kil s CT3 sI A33 1 B3 dm
Hence, according to Eq. (4.61), the system is shown in Fig. 4.11.
142
Fig. 4.11 Closed-loop
system installed with the
SSSC stabilizer
K(s)
0
s
Electromechanical oscillation loop
1
s
K c (s)
K o (s)
K il (s)
+
y
Tsssc (s)
From Fig. 4.11, it can be seen that with the SSSC stabilizer installed, the electric
torque provided by the SSSC stabilizer to the electromechanical oscillation loop of
generator is
DTet
Kc jxs Ko jxs Tsssc jxs

Dx
1 Kil jxs Tsssc jxs
4:62
where xs is the angular frequency of power system oscillation. An ideal SSSC

stabilizer should contribute a pure positive damping torque to the electromechanical
oscillation loop, DTdt Dsssc Dx, that is
Kc jxs Ko jxs Tsssc jxs
Dsssc
1 Kil jxs Tsssc jxs
4:63
Hence,
Dsssc Kc jxs Ko jxs Dsssc Kil jxs Tsssc jxs
Fsssc jxs Tsssc jxs
4:64
According to Eq. (4.64), the phase compensation method can be used to design the
SSSC stabilizer by setting
h u;
Tsssc
where Tsssc jxs Tsssc \hFsssc jxs Fsssc \u.
Dsssc
Fsssc
4:65
4.3 Damping Function of a Unied Power Flow Controller
4.3
143
Damping Function of a Unied Power Flow Controller

(UPFC) Installed in a Single-Machine Innite-Bus
Power System
4.3.1
Mathematical Model of a Single-Machine Innite-Bus

Power System Installed with a UPFC
4.3.1.1
Dynamic Model of a UPFC
Figure 4.12 shows the conguration of a single-machine innite-bus power system

installed with a unied power flow controller (UPFC). The UPFC consists of an
excitation transformer (ET), a boosting transformer (BT), two three-phase
GTO-based voltage source converters (VSCs), and a DC link capacitor. In
Fig. 4.12, me ; mb and de ; db are the amplitude modulation ratio and phase angle of
the pulse width modulation (PWM) implemented by the shunt and series VSC,
respectively[5].
In Fig. 4.12, the shunt and series synchronous voltage source (SVS) can be
denoted as
Ves me ke Vdc \wes
4:66
Vbs mb kb Vdc \wbs
Vt
Vb
x tb
Itb
Pt
I te
Vet
Vbs
I bs
x te
Pte
Vbt
x bs
Ies
VSC B
VSC E
x es
BT
ET
Iedc
I bdc
Cdc
Ves
m e e
UPFC
m b b
Fig. 4.12 A single-machine innite-bus power system installed with a UPFC
x bt
144
y
Vet
Vbt
jx es Ies
e
Ves
jx bs I bs
es
Vbs
bs
Fig. 4.13 Phase diagram of shunt and series SVS
where wes and wbs are the angle between d-axis of dq coordinate of the synchronous
generator and voltage Ves and Vbs , respectively, as shown in Fig. 4.13, which gives
Vet jxes Ies Ves
Vbt jxbs Ibs Vbs
4:67
The active power injected into the DC side of the UPFC from the shunt and series
transformer is the active power consumed by the shunt and series SVS, respec

tively, Pes ReVes Ies and Pbs ReVbs Ibs , where Re and * denote the real
part and conjugate of a complex variable, respectively. From Eqs. (4.66) and (4.67),
it can be obtained that
Ves Vet
me ke Vdc Vet
sin de
sin de
xes
xe
Vbs Vbt
mb kb Vdc Vbt
Pbs
sin db
sin db
xbs
xbs
Pes
4:68
The shunt and series active power injection should be equal to that received at the
DC capacitor from the shunt and series VSC, respectively, that is
me ke Vdc Vet
sin de Iedc Vdc
xes
mb kb Vdc Vbt
Pbs
sin db Ibdc Vdc
xbs
Pes
4:69
145

me ke Vet
sin de
xes
mb kb Vbt
sin db
xbs
Iedc
Ibdc
4:70
Alternatively, in the dq coordinate as shown in Fig. 4.13,

Pes vesd iesd vesq iesq iesd me ke Vdc cos wes iesq me ke Vdc sin wes Iedc Vdc
Pbs vbsd ibsd vbsq ibsq ibsd mb kb Vdc cos wbs ibsq mb kb Vdc sin wbs Ibdc Vdc
4:71
Iedc iesd me ke cos wes iesq me ke sin wes
Ibdc ibsd mb kb cos wbs ibsq mb kb sin wbs
4:72
Hence, from Eq. (4.70), dynamic equation of the UPFC can be obtained as

Iedc Ibdc
1 me ke Vet
mb kb Vbt
V_ dc
sin de
sin db
Cdc
Cdc
xes
xbs
4:73
Alternatively, from Eq. (4.72), it can have

1
V_ dc
iesd me ke cos wes iesq me ke sin wes ibsd mb kb cos wbs ibsq mb kb sin wbs
Cdc
4:74
where
vetq
de
vetd
vbtq
wbs tan1
db
vbtd
wes tan1
4:75
146
4.3.1.2
Nonlinear Model of the Single-Machine Innite-Bus Power

System Installed with the UPFC

Vt jxtb Itb Vb
Vt jxte Ite Vet
4:76
Vet Vbt jxbt Ibs Vb

which can be expressed in the dq coordinate as
vtd jvtq jxtb itbd jitbq Vb sin d jVb cos d
xtb itbq Vb sin d jxtb itbd jVb cos d
vtd jvtq jxte iesd ibsd jiesq jibsq vetd jvetq
4:77
vetd jvetq vbtd jvbtq jxbt ibsd xbt ibsq Vb sin d jVb cos d
From the rst equation in Eq. (4.77), it can be obtained that
vtq Vb cos d
xtb
Vb sin d vtd
xtb
itbd
itbq
4:78
Because
vtd xq iteq itbq
vtq E0q x0d ited itbd
4:79

vtd
vtq
xq xtb iteq xq Vb sin d

xq xtb
0
xtb Eq x0d xtb ited x0d Vb cos d
x0d xtb
4:80
147
The second and third equations in Eq. (4.77) give

vtd jvtq jxte iesd ibsd jiesq jibsq xes iesq vesd jxes iesd jvesq
xes iesq vesd jxes iesd jvesq
xbs ibsq vbsd jxbs ibsd jvbsq jxbt ibsd xbt ibsq Vb sin d jVb cos d
That is
vtd xte iesq ibsq xes iesq vesd
vtq xte iesd ibsd xes iesd vesq
4:81
xes iesq vesd xbs ibsq vbsd xbt ibsq Vb sin d

xes iesd vesq xbs ibsd vbsq xbt ibsd Vb cos d
xq xtb xte xq xtb xes xq xtb iesq
xq xtb xte xq xtb ibsq xq xtb vesd xq Vb sin d
xte x0d xtb xes x0d xtb x0d xte iesd
xte x0d xtb x0d xtb ibsd xtb E0q x0d Vb cos d x0d xtb vesq
xes iesq xbs xbt ibsq vbsd vesd Vb sin d
xes iesd xbs xbt ibsd vbsq vesq Vb cos d
which in the matrix form is
iesq
xq xtb xte xes xq xtb
xq xtb xte xq xtb
1
xes
xes xbt

xq xtb me ke Vdc cos wes xq Vb sin d

mb kb Vdc cos wbs me ke Vdc cos wes Vb sin d

xte xes x0d xtb x0d xte xte x0d xtb x0d xtb 1
iesd
ibsd
xes
xbs xbt

xtb E0q x0d Vb cos d x0d xtb me ke Vdc sin wes

mb kb Vdc sin wbs me ke Vdc sin wes Vb cos d
ibsq
4:82
148
For the single-machine innite-bus power system with the UPFC installed, it can
have
ited iesd ibsd
iteq iesq ibsq
Pt vtq iteq itbq vtd ited itbd
Eq E0q xd x0d ited itbd
q
Vt v2td v2tq
4:83
Equations (4.74), (4.75), (4.78), (4.80), (4.82), and (4.83), together with the differential equations of synchronous generator of Eq. (2.35), form the nonlinear
mathematical model of the single-machine innite-bus power system with the
UPFC installed.
4.3.1.3
Linearized Model of the Single-Machine Innite-Bus Power

System Installed with the UPFC

Diesq
Dibsq
3
3
2
2
3
2
3
Dme
Dme

Dd
Dd
7
6 Dwes 7 Diesd
6
0 6 Dwes 7
04
0 5
7
F0q 4 DE0q 5 G0q 6
4 Dmb 5; Dibsd Fd DEq Gd 4 Dmb 5
DVdc
DVdc
Dwbs
Dwbs
2
4:84
where

f q11 0 f q13
xte 2xq xtb
1 xbs xbt
xR1
xes
xq xte xte xes xq xtb f q21 0 f q23

xte xtb
f d11 xL f d13
1 xbs xbt
0
Fd
xR2
xes
xte xes x0d xtb x0d xte f d21 0 f d23
#

"
g
0
0
g
x
x
x
2x
1
bs
bt
te
q
tb
q11
q12
G0q
xR1
xes
xq xte xte xes xq xtb gq21 gq22 gq23 gq24

xte xtb
gd11 gd12
0
0
1 xbs xbt
0
Gd
xR2
xes
xte xes x0d xtb x0d xte gd21 gd22 gd23 gd24
F0q
xR1 xq xte xte xes xq xtb xbs xbt x2es 2xq xtb
xR2 xte xes x0d xtb x0d xte xbs xbt x2te xtb
f q11 xq xtb me0 ke Vdc0 sin wes0 xq Vb cos d0

f q13 xq xtb me0 ke cos wes0
f q21 mb0 kb Vdc0 sin wbs0 me0 ke Vdc0 sin wes0 Vb cos d0
f q23 mb0 kb cos wbs0 me0 ke cos wes0
f d11 x0d xtb me0 ke Vdc0 cos wes0 x0d Vb sin d0
f d13 x0d xtb me0 ke sin wes0
f d21 me0 ke Vdc cos wes0 mb0 kb Vdc0 cos wbs0 Vb sin d0
f d23 mb0 kb sin wbs0 me0 ke sin wes0
gq11 xq xtb ke Vdc0 cos wes0
gd12 xq xtb me0 kE Vdc0 sin wes0
gq21 ke Vdc0 cos wes0
gq22 me0 ke Vdc0 sin wes0
gq23 kb Vdc0 cos wbs0
gq24 mb0 kb Vdc0 sin wbs0
gd11 x0d xtb ke Vdc0 sin wes0
gd12 x0d xtb me0 ke Vdc0 cos wes0

gd21 ke Vdc0 sin we0
gd22 me0 ke Vdc cos wes0
gd23 kb Vdc0 sin wbs0
gd24 mb0 kb Vdc0 cos wbs0
vetd jvetq jxes iesd jiesq me ke Vdc cos wes jme ke Vdc sin wes
vbtd jvbtq jxbs ibsd jibsq mb kb Vdc cos wbs jmb kb Vdc sin wbs
The above equations give
vesd xes iesq me ke Vdc cos wes
vesq xes iesd me ke Vdc sin wes
vbsd xbs ibsq mb kb Vdc cos wbs
vbsq xbs ibsd mb kb Vdc sin wbs
149
150
By using Eq. (4.84), linearization of the above equation can be obtained as

Dvesd xes f 0q11 Dd xE f 0q12 DE0q xes f 0q13 me0 ke cos wes0 DVdc
xes g0q11 ke Vdc0 cos wes0 Dme xes g0q12 me0 ke Vdc0 sin wes0 Dwes
xes g0q13 Dmb xes g0q14 Dwbs
aed1 Dd aed2 DE0q aed3 DVdc aed4 Dme aed5 Dwes aed6 Dmb aed7 Dwbs
Dvesq xes f 0d11 Dd xes f 0d12 DE0q xes f 0d13 me0 ke sin wes0 DVdc
xes g0d11 ke Vdc0 sin wes0 Dme
xes g0d12 me0 ke Vdc0 cos wes0 Dwes xes g0d13 Dmb xes g0d14 Dwbs
aeq1 Dd aeq2 DE0q aeq3 DVdc aeq4 Dme aeq5 Dwes aeq6 Dmb aeq7 Dwbs
Dvbsd xbs f 0q21 Dd f 0q22 DE0q xbs f 0q23 mb0 kb cos wbs0 DVdc
xbs g0q23 kb Vdc0 cos wbs0 Dmb xbs g0q24 mb0 kb Vdc0 sin wbs0 Dwbs
xbs g0q21 Dme xbs g0q22 Dwes
abd1 Dd abd2 DE0q abd3 DVdc abd4 Dme abd5 Dwes abd6 Dmb abd7 Dwbs
Dvbsq xbs f 0d21 Dd xbs f 0d22 DE0q xbs f 0d23 mb0 kb sin wbs0 DVdc
xbs g0d23 kb Vdc0 sin wbs0 Dmb xbs g0d24 mb0 kb Vdc0 cos wbs0 Dwbs
xbs g0d21 Dme xbs g0d22 Dwes

abq1 Dd abq2 DE0q abq3 DVdc abq4 Dme abq5 Dwes abq6 Dmb abq7 Dwbs
By using the above equations, linearization of Eq. (4.75) can be obtained as

Dwes a0e1 Dd a0e2 DE0q a0e3 DVdc a0e4 Dme
a0e5 Dwes a0e6 Dmb a0e7 Dwbs Dde
Dwbs a0b1 Dd a0b2 DE0q a0b3 DVdc a0b4 Dme
a0b5 Dwes a0b6 Dmb a0b7 Dwbs Ddb

where
1
vesd0 aeqi vesq0 aedi
v2esd0 v2esq0
1
a0bi 2
vbsd0 abqi vbsq0 abdi
vbsd0 v2bsq0
a0ei
i 1; 2; 3; 4; 5; 6; 7
Rearrangement of above equations is

Dwes
Dwbs
1 a0e5
a0b5
a0e7
1 a0b7
a0e4
a0e6
a0b4
a0b6

3
1
Dme
Dmb
a0e1
a0b1
3
2
Dd
a0e3 6
7
4 DE0q 5
a0b3
DVdc

!
Dde
0
a0e2
a0b2
1
0
151
1
Ddb
Dd

Dme
Dde
6
0 7
AEB1 4 DEq 5 AEB2
AEB3
Dmb
Ddb
DVdc
4:85

3
Dme
6 Dd 7
Diesq
6 e7
7
6
Fq 4 DE0q 5 Gq 6
7
4 Dmb 5
Dibsq
DVdc
Ddb
3
2
3
2
Dme
Dd

6 Dd 7
Diesd
6 e7
7
6
Fd 4 DE0q 5 Gd 6
7
4 Dmb 5
Dibsd
DVdc
Ddb
Dd
4:86

xq xtb
xq Vb sin d0
Dvtd
Dd
xq xtb
xq xtb

xq xtb
Diesq
xq xtb Dibsq

4:87
0
Diesd
x0d xtb x0d xtb
xd Vb sin d0
xtb
0
Dvtq
Dd 0
DE 0
xd xtb x0d xtb Dibsd
x0d xtb
xd xtb q
Substituting Eq. (4.86) into (4.87) gives

2
3
3
Dme
Dd
6 Dd 7
6 e7
6
7
Dvtd Hd 4 DE0q 5 Ld 6
7
4 Dmb 5
DVdc
Ddb
2
3
2
3
Dme
Dd
6 Dd 7
6 e7
6
7
Dvtq Hq 4 DE0q 5 Lq 6
7
4 Dmb 5
DVdc
Ddb
2
4:88
152
where

xq xtb
xq xtb
xq Vb sin d0
0 0
F
xq xtb xq xtb q
xq xtb

xq xtb
xq xtb
Gq
xq xtb xq xtb
0

xd Vb sin d0
xtb
x0d xtb x0d xtb
0 0
Fd
x0d xtb
x0d xtb
xd xtb x0d xtb

x0 xtb x0d xtb
0 d
Gd
xd xtb x0d xtb
Hd
Ld
Hq
Lq

Vb sin d0
1
Dd
Dvtq
xtb
xtb
Vb cos d0
1
Dd Dvtd
xtb
xtb
Ditbd
Ditbq
4:89

Dited Diesd Dibsd
Diteq Diesq Dibsq
DPt iteq0 itbq0 Dvtq ited0 itbd0 Dvtd
vtq0 Diteq Ditbq vtd0 Dited Ditbd
DEq DE0q xd x0d Dited Ditbd
vtq0
vtd0
Dvtd
Dvtq
DVt
Vt0
Vt0
4:90
By using Eqs. (4.86), (4.88), and (4.89) and the rst two equations in Eq. (4.90), the
last three equations in Eq. (4.90) become
DPt iteq0 itbq0 Dvtq ited0 itbd0 Dvtd
vtq0 Diesq Dibsq Ditbq vtd0 Diesd Dibsd Dibsd

Vb cos d0
Vb sin d0
vtd0
Dd
vtq0
xtb
xtb

vtq0
vtd0
ited0 itbd0
Dvtd iteq0 itbq0
Dvtq
xtb
xtb

Diesq
Diesd
vtd0 vtd0
vtq0 vtq0
Diesq
Diesd
3
2
3
2
Dme
Dd
6 Dd 7
6 e7
7
6
K1 K2 Kpdc 4 DE0q 5 Kpme Kpde Kpmb Kpdb 6
7
4 Dmb 5
DVdc
Ddb
4:91
DE0q
x0d
x0d
Diesd
153
xd
xd
Dibsd

V
sin
d
1
b
0
0
xd xd
Dd
Dvtq
xtb
xtb
Vb sin d0
1
xd x0d
Dd DE0q xd x0d Dvtq
xtb
xtb

Diesd
0
0
xd xd xd xd
Dibsd
2
3
Dd
6
7
K4 K3 Kqdc 4 DE0q 5 Kqme Kqde Kqme
DEq
DVdc
Dme
6 Dd 7
6 e7
Kqde 6
7
4 Dmb 5
Ddb
4:92
DVt K5
K6
3
Dd
Kvdc 4 DE0q 5 Kvme
DVdc
3
Dme
6 Dde 7
7
Kvdb 6
4 Dmb 5 4:93
Ddb
2
Kvde
Kvmb
where
K1
K2
h
i
Kpdc vtq0 Vbxcos d2 vted0 Vbxsin d0 0 0
tb
tb

vteq0
vted0
Hd iteq0 itbq0
Hq
ited0 itbd0
xtb
xtb
vtq0
Kpme
Kpde
K4
Kqme
Kqde
K5
Kqme
Kqde
Kpmb
K3
Kqmb
K6
Kqmb
vtq0 Fq vtd0
vtd0 Fd
vteq0
Ld
xtb
vted0
iteq0 itbq0
Lq vtq0 vtq0 Gq vtd0
xtb
h
i
Kqdc xd x0d Vbxsin d0 1 0
tb
xd x0d
Hq xd x0d xd x0d Fd
xtb
xd x0d
Kqdb
Lq xd x0d xd x0d Gd
xtb
vtq0
vtd0
Hq
Kvdc
Hd
Vt0
E
vtq0
vtd0
Kqdb
Ld
Lq
Vt0
Vt0
Kpdb ited0 itbd0
vtd0 Gd
154

Cdc DV_ dc me0 ke cos wes0
me0 ke sin wes0

Diesd
mb0 kb cos wbs0
Dibsd

Diesq
mb0 kb sin wbs0
Dibsq
ke iesd0 cos wes0 Dme ke iesq0 sin wes0 Dme kb ibsd0 cos wbs0 Dmb
kb ibsq0 sin wbs0 Dmb me0 ke iesd0 sin wes0 Dd Dde
me0 ke iesq0 cos wes0 Dd Dde mb0 kb ibsd0 sin wbs0 Dd Ddb
mb0 kb ibsq0 cos wbs0 Dd Ddb
me0 ke iesd0 sin wes0 me0 ke iesq0 cos wes0 mb0 kb ibsd0 sin wbs0
mb0 kb ibsq0 cos wbs0 Dd me0 ke cos wes0
me0 ke sin wes0

Diesq
mb0 kb sin wbs0
Dibsq

Diesd
mb0 kb cos wbs0
DiBbsd
ke iesd0 cos wes0 ke iesq0 sin wes0 Dme

kb ibsd0 cos wbs0 kb ibsq0 sin wbs0 Dmb
me0 ke iesq0 cos wes0 me0 ke iesd0 sin wes0 Dde
mb0 kb ibsq0 cos wbs0 mb0 kb ibsd0 sin wbs0 Ddb Ddb
3
2
2
3
Dme
Dd
6 Dd 7
6 e7
6
7
K7 K8 K9 4 DE0q 5 Kdme Kdde Kdmb Kddb 6
7
4 Dmb 5
DVdc
Ddb
4:94
where
K7
K8 K9
3T
2
me0 ke iesd0 sin wes0 me0 ke iesq0 cos wes0 mb0 kb ibsd0 sin wbs0 mb0 kb ibsq0 cos wbs0
7
6
0
4
5
0
me0 ke cos wes0
mb0 kb cos wB0 Fd me0 ke sin wE0

2
Kdme
Kdde
Kdmb
mb0 kb sin wB0 Fq
ke iesd0 cos wes0 ke iesq0 sin wes0
6 me0 ke iesq0 cos w me0 ke iesd0 sin w

es0
es0
6
Kddb 6
4
kb ibsd0 cos wbs0 kb ibsq0 sin wbs0
3T
7
7
7
5
mb0 kb ibsq0 cos wbs0 mb0 kb ibsd0 sin wbs0

me0 ke cos wes0
mb0 kb cos wbs0 Gd me0 ke sin wes0
mb0 kb sin wbs0 Gq
155
By substituting Eqs. (4.91)(4.93) into Eq. (2.37) (without the PSS control) and
arranging the equations obtained together with Eq. (4.94) in matrix form, the
extended HeffronPhillips model of the single-machine innite-bus power system
with the UPFC installed can be obtained as
3
2
0
xo
0
0
0
3
3
2
6
Dd_
Kpdc 7
K2
D
7 Dd
6 K1

0

6 _ 7 6
M
M
M
M 76 Dx 7
7
6 Dx 7 6
6
Kqdc 7
K
K
7
7
6
6
6
1
3
4
0
0
0 7
0 7
0
7
6 DE_ 0 7 6 T0
6
DE
Td0
Td0
Td0 76
q 7
q 7
d0
6
6
76
7 6
7
6
4 DE_ 0fd 5 6 KA K5
4 DE0fd 5
0
KTA K6 T1 KATKvdc 7
7
6
T
A
A
A
A
5 DVdc
4
DV_ dc
K8
K9
K7
0
0
Cdc
Cdc
Cdc
3
2
0
0
0
0
K
K
Kpmb
6 Kpme
3
2
Mpde
M
Mpdb 7
7 Dme
6
M
7
6
76 Dd 7
6
K
K
K
K
qde
qmB
qdB
76 e 7
6 qme
6
T0d0
T0d0
T0d0
T0d0 76
7
74
6
5
6 KA Kvme KA Kvde KA Kvmb KA Kvdb 7 Dmb
7
6
TA
TA
TA
TA 5 Ddb
4
Kdde
Kddb
Kdme
Kdmb
Cdc
Cdc
Cdc
Cdc
2
4:95
The linearized model is shown in Figs. 4.14 and 4.15, where

kp Kpdc
kq Kqdc
Kpme
Kpde
Kpmb
Kpdb
Kqme
Kqde
Kqmb
Kqdb
kv Kvdc Kvme
kd Kdme Kdde
Kvde Kvmb Kvdb

Kdmb Kddb
uupfc1 Dme Dde Dmb

T
uupfc DVdc uTupfc1
4:96
Ddb T
4.3.2
Design of a UPFC Stabilizer Installed

in a Single-Machine Innite-Bus Power System
4.3.2.1
Selection of Modulation Signal to Add the Damping Control

Signal of the UPFC Stabilizer by Using the Controllability
and Observability Index [6, 7]
Damping control signal of a UPFC stabilizer can be superimposed on a modulation

signal of the PWM control, Dme , Dde , Dmb , or Ddb . Selection of the modulation
signal to add the damping control signal ought to ensure the most effectiveness of
156
K1
1
sM + D
0
s
K4
K5
K2
kp
uupfc
kv
kq
E q '
1
Td0 's + K 3
KA
1 + sTA
K6
Fig. 4.14 Extended HeffronPhillips model of power system with a UPFC installedpart of the
power system
Fig. 4.15 Extended Heffron

Phillips model of power
system with a UPFC installed
part of the UPFC
K7
E q '
K8
1
sCdc K 9
Vdc
kd
uupfc1
uupfc1
UPFC stabilizer. It can be carried out by use of the modal controllability and
observability index introduced in Sect. 2.2.1.2.
Linearized model of the single-machine innite-bus power system installed with
the UPFC of Eq. (4.95) can be written as
X_ AX
4
X
157
bk Duk
4:97
k1
where
2
0
xo
6
Dd
K
D
6 1
M
7
6
6
M
6
6 Dx 7
7
6 K0 4
6
0
0 7
6
X6
Td0
6 DEq 7; A 6
6 K K
6
0 7
6 A 5
4 DEfd 5
0
6
TA
4
DVdc
K7
0
Cdc
3
3
2
2
0
0
6 Kpme 7
6 Kpde 7
6
6
M 7
M 7
7
7
6
6
6
6
Kqde 7
Kqme 7
6 0 7
6 0 7
b1 6
Td0 7; b3
Td0 7; b2 6
7
7
6
6
6 KA Kvme 7
6 KA Kvde 7
7
6
6
TA 5
TA 7
5
4
4
Kdme
Kdde
Cdc
Cdc
Du1 Dme ; Du2 Dde ; Du3 Dmb ;
2
3
0
7
K
0
Mpdc 7
7
Kqdc 7
1
0 7
T0d0
Td0 7
7;
7
K
1
A Kvdc 7
T T 7
A
A
5
K9
0
Cdc
3
3
2
0
0
6 Kpmb 7
6 Kpdb 7
6
6
M 7
M 7
7
7
6
6
6
6
Kqdb 7
Kqmb 7
6 0 7
6 0 7
6
Td0 7;
Td0 7; b4 6
7
7
6
6
6 KA Kvmb 7
6 KA Kvdb 7
7
6
6
TA 5
TA 7
5
4
4
Kdmb
Kddb
Cdc
Cdc
Du4 Ddb
0
2
K
M
K0 3
Td0
K
TA K6
A
K8
Cdc
2
If the transfer function and feedback signal of the UPFC stabilizer are Tupfck s) and
yk , respectively, state-space representation of the power system with the UPFC
stabilizer is
X_ AX
4
X
bk Duk
k1
yk ckT X
4:98
Duk Tupfck syk

If ki is the electromechanical oscillation mode of the power system, from Eqs. (2.50
) and (2.54), it can have
Avi ki vi
wTi A wTi ki
4:99
where vi and wTi are the right eigenvector and left eigenvector of state matrix, A,
with respect to mode ki , respectively. From the state transformation of Eqs. (2.53),
(2.55), (2.56), and (4.98), it can have
158
Z_ V1 AVZ
4
X
V1 bk Duk K Z
k1
yk
4
X
V1 bk Duk
4:100
k1
cTk VZ
Duk Tupfck syk

where
2
V v1
v2
. . . vn ;
k1
60
6
K6
40
0
0
k2
0
0
0
0
..
.
0
3
0
07
7
7
05
kn
According to Eq. (2.53),

V1 w1
w2
. . . w M T W T
4:101

4
X
1
wTj bk Duk
s kj
k1
X
yk
cTk vj zj ; j 1; 2; . . .
zj
4:102
Duk Tupfck syk

The system represented by Eq. (4.102) is shown in Fig. 4.16.
From Fig. 4.16, it can be seen that wiT bk measures the influence of the UPFC
stabilizer on the oscillation mode and ckT vi measures how much the oscillation
mode is observed in the feedback signal of the stabilizer. Hence, it is dened that as
Fig. 4.16 Modal
decomposition of the system
to be installed with UPFC
stabilizer
u k
b k1
1
s 1
Ck1
bk 2
1
s 2
Ck 2
b kn
1
s n
Ckn
Tupfck (s)
yk
159
far as the oscillation mode ki is concerned, the controllability index and observability index of the UPFC stabilizer are, respectively,
bik wiT bk ;
cik ckT vi
4:103
Their product is the residue

Rik bik cik
4:104
which measures the overall influence of the stabilizer on the oscillation mode.
Obviously, the controllability index, observability index, and residue are calculated from the open-loop system model as far as the UPFC stabilizer is concerned. Hence, it can be used to predict the effectiveness of the stabilizer to be
installed. The controllability index can be used to select the most effective modulation signal to be superimposed with the damping control signal to enhance the
damping of system electromechanical oscillation mode.
4.3.2.2
Selection by Damping Torque Calculation
Electric torque contribution from a UPFC stabilizer can be calculated from the
forward path of UPFC damping control signal. The forward path can be obtained
from Figs. 4.14 and 4.15 with the blocks associated with K4 , K5 , and K7 ignored as
shown in Fig. 4.17,
Du1 Dme ;
Du2 Dde ;
Du3 Dmb ;
Du4 Ddb
kpuk
when k 1; kpuk Kdme ;

Kpdc Kpme ; kquk Kqdc
Kqme ;
kpuk
when k 2; kpuk Kdde ;

Kpdc Kpde ; kquk Kqdc
Kqde ;
kpuk
kpuk
kvuk Kvdc
kvuk Kvdc
Kvme
Kvde
when k 3; kpuk Kdmb ;

Kpdc Kpmb ; kquk Kqdc Kqmb ; kvuk Kvdc Kvmb
when k 4; kpuk Kddb ;
Kpdc Kpdb ; kquk Kqdc Kqdb ; kvuk Kvdc Kvdb
The forward path is described by the following state-space representation

X_ 1 A1 X1 b1k Duk
DTuk cTupfc1 X1 kpuk Duk ;
k 1; 2; 3; 4
4:105
160
Fig. 4.17 Forward path of
UPFC stabilizer
Tuk
K8
u k
K duk
1
sCdc K 9
K2
k puk
Vdc
u
k
k vuk
k quk
E q '
1
Td0 's + K 3
KA
1 + sTA
K6
where
Du1 Dme ;
Du2 Dde ; Du3 Dmb ; Du4 Ddb

3
2
K
1
3
2
qdc
K0 3
0
0
0
Td0
Td0
Td0 7
DEq
6
7
6
7
6
K
K
1
A K6
A Kvdc 7
X1 4 DE0fd 5; A1 6
T T 7
6 TA
A
A
5
4
DVdc
K8
K
9
0
Cdc
Cdc
T
T
b11 Kqme Kvme Kdme ; b12 Kqde Kvde Kdde ;
b13 Kqmb
cupfc1 K2
Kvmb
0
Kdmb T ;
b14 Kqdb
Kvdb
Kddb T ;
Kpdc T
From Eq. (4.105), transfer function of the forward path can be obtained as
T
Fupfck s cupfc1
sI A1 1 b1k ;
k 1; 2; 3; 4
4:106
Hence, the electric torque contributed from the UPFC stabilizer is

DTuk Fupfck jxs Tupfck jxs yk ;
k 1; 2; 3; 4
4:107
where xs is the angular frequency of power oscillation. From Eq. (4.107), it can be
seen that the following index can also be used to predict the effectiveness of the
UPFC stabilizer to be installed. The index can be used to select the modulation
signal to add the damping control signal of the UPFC stabilizer

Fupfck jxs ;
k 1; 2; 3; 4
4:108
4.3.2.3
161
Robustness of the Damping Control to the Variations of Power

System Operating Conditions
Discussions so far have demonstrated that the effectiveness of a stabilizer may

change with variations of power system operating conditions (load conditions and
network congurations). Design of the stabilizer is carried out based on the linearized model which is obtained at a selected operating condition. Hence, the
effectiveness of stabilizer is only guaranteed at the selected operating condition.
However, operating conditions of a power system are variable. In order to ensure
that the stabilizer is effective over a certain range of power system operating
conditions, when the stabilizer is designed its robustness to the variations of system
operating conditions should be considered. This is the topic to be discussed in this
section.
It is assumed that the set of system known operating conditions is X0 and the
stabilizer is to be designed to ensure its robustness over X0 , that is, to guarantee the
effectiveness of the stabilizer in damping power system oscillations when system
operating conditions change within X0 . If it is known that at an operating condition
rr 2 X0 , the stabilizer is least effective, rr 2 X0 can be selected as the operating
condition at which the stabilizer is designed. By doing so, the design can ensure the
effectiveness of stabilizer over the set of system operating conditions and hence
guarantees the robustness of the stabilizer to the variations of system operating
conditions within set X0 . Based on this principle, two examples to design a robust
stabilizer are introduced as follows.
1. Robust selection of modulation signal to add UPFC stabilizer
Selection of a modulation signal to add the damping control signal of a UPFC
stabilizer is in order to maximize its effectiveness. When the controllability index is
used, the criterion of selection is
Duselected Maxbik uTi bk ;
Duk
Duk 2 fDme ; Dmb ; Dde ; Ddb g
4:109
When the damping torque index is used, according to Eq. (4.108), the criterion of
selection is

Duselected Max Fupfck jxs ; k 1; 2; 3; 4 ; Duk 2 fDme ; Dmb ; Dde ; Ddb g
Duk
4:110
To consider the robustness of stabilizer to the variations of system operating conditions, the above criteria of Eqs. (4.109) and (4.110) should be calculated at a
selected operating condition of the power system when
rselected Minbik uTi bk ;
r
r 2 X0
4:111
162
or

rselected Min Fupfck jxs ; k 1; 2; 3; 4 ;
r
r 2 X0
4:112
Equations (4.109)(4.112) mean that the selection is made at an operating condition

of power system where the stabilizer is predicted to be least effective and the most
effective modulation signal is selected to add the UPFC stabilizer at the selected
operating condition.
2. Design of robust stabilizer by the phase compensation method
Denote the transfer function of a stabilizer installed in a single-machine innite-bus
power system to be Tstab s Kstab Tstabk s, where Kstab is the gain of the stabilizer.
The feedback signal of the stabilizer is ystab , and its decomposition into the damping
torque and synchronizing torque is

x0 Ksstab
ystab Kdstab Dx Ksstab Dd Kdstab
Dx Ky jxs Dx
jxs
4:113
If the forward path of stabilizer to the electromechanical oscillation loop of generator is Fstab s), the damping torque provided by the stabilizer is
DTDstab ReFstab jxs Ky jxs Tstab jxs Dx
4:114
and the synchronizing torque is

DTSstab ImFstab jxs Ky jxs Tstab jxs
xs
Dd
x0
4:115
From Eq. (4.114), it can be seen that with the variations of system operating
conditions, change of the damping torque contributed by the stabilizer is mainly
determined by variations of Fstab jxs Ky jxs . At an operating condition rj 2 X0 , if
it is denoted that
Fstab jxs Ky jxs Hj \uj ;
at rj 2 X0
Tstab jxs H\ /
4:116
4:117
From Eqs. (4.114) and (4.115), the damping torque and synchronizing torque
provided by the stabilizer at rj 2 X0 can be obtained as
DTdstab Hj H cosuj /Dx
DTsstab Hj H sinuj /
xs
Dd
x0
4:118
4:119
163
If it has
umax maxuj ;
rj 2 X0
4:120
in order for the stabilizer to provide the power system with positive synchronizing
torque at all known operating conditions rj 2 X0 , design of the stabilizer must
ensure
/
uj ; rj 2 X0
4:120a
/ umax
4:121
Hence, simply it can choose
Therefore, from Eqs. (4.118) and (4.121), the damping torque contributed by the
stabilizer can be obtained as
DTdstab Hj H cosuj umax Dx
4:122
Equation (4.122) indicates that the effectiveness of stabilizer to be designed is

measured by Hj cosuj umax at rj 2 X0 . Hence, the operating condition at which
the stabilizer will be least effective can be selected by using the following criterion:
Hselected cosuselected umax MinHj cosuj umax ; rj 2 X0
j
4:123
At rselected 2 X0 determined by using the criterion of Eq. (4.123), the stabilizer can
be designed. The design will ensure that more damping torque will be provided by
the stabilizer when the power system operates at other operating conditions.
Therefore, the design guarantees the robustness of the stabilizer to the variations of
power system operating conditions.
4.4
Examples
4.4.1
An Example Single-Machine Innite-Bus Power

System Installed with a BESS Stabilizer
4.4.1.1
Extended HeffronPhillips Model of the Example Power

System Installed with a BESS
Conguration of an example single-machine innite-bus power system installed

with a shunt-connected BESS is as same as that shown in Fig. 4.2. The model of
BESS is shown in Fig. 4.18.
164
rbess
Vc
VSC
Cdc
Vdc
Vbess
BESS
Fig. 4.18 Model of BESS
Idc2
Vbess Vdc
rbess
4:124
Parameters of the example system are as follows:

Generator
The AVR
Transmission line
The BESS
Steady-state operating point
xd 1:0; xq 0:6; x0d 0:3; M 8s:; D 0; T0d0 5:044s:

KA 100; TA 0:01
xts 0:3; xsb 0:3; xs 0:15
k 3; Cdc 1:0; Vdc0 1:0; Vbess 1:0; rbess 0:01
Pt0 0:5; Vt0 1:0; Vb 1:0; Vs0 1:0
Computation of initial values of various variables at the steady-state operation of

example power system is as same as that given in Sect. 3.3.1.2 as follows:
Qb0
2s
3

2
Vb 4
x
P
sb
t0
V2s0
Vb 5 0:0337
Xsb
Vb
Pt0 jQb0
0:5 j0:0377
Vb
Vs0 jxsb Isb Vb 0:9887 j0:15 1:0\8:6
q
V2t0 V2s0 Pt0 xts 2 V2s0
Qs0
0:0377
xts
Pt0 jQs0
Its0
0:4887 j0:1123
Vs0
Is0 Its0 Isb0 0:0113 j0:0746
Isb0
Vt0 jxts Its0 Vs0 0:955 j0:2966

EQ Vt0 jxq Its0 0:8876 j0:5898 1:07\33:6
4.4 Examples
165
Denote the real and imaginary part of a complex number by Re( ) and Im( ),
respectively, from the phasor diagram of Fig. 4.3 it can have
itsd0 ReIts0 sin d0 ImIts0 cos d0 0:1769
itsq0 ReIts0 cos d0 ImIts0 sin d0 0:4692
vtd0 ReVt0 sin d0 ImVt0 cos d0 0:2815
vtq0 ReVt0 cos d0 ImVt0 sin d0 0:9596
isd0 ReIs0 sin d0 ImIs0 cos d0 0:0684
isq0 ReIs0 cos d0 ImIs0 sin d0 0:0318
E0q0 EQ xd x0d itsd0 1:0126
Vc0 Vs0 jxs Is0 0:9999 j0:1517 1:011\8:6
Vc0
m0
0:3371
kVdc0
From Fig. 4.3, it can be seen that w0 and wx0 8:6 is the phase of Vc0 in the d
q coordinate and xy coordinate, respectively. Hence,
w0 wx0
p
d0 65:0250
2
From Eqs. (4.7) and (4.15), it is obtained that

c11
c21
c12
c22
0:6667
2:6667

1:0000
;
0:6667
d11
d21
d12
d22
0:9524
2:8571
1:4286
0:9524
By using Eqs. (4.16)(4.19), following coefcients in the HeffronPhillips

model of Figs. 4.4 and 4.5 are obtained
K1 0:3330; K2 0:6702; K3 2; K4 0:1845; K5 0:0290;
K6 0:5483; K7 1:9772; K8 0:4067; K9 0:0746
Kdm 0:2212; Kpm 0:5355; Kqm 1:8130; Kpw 0:7086;
Kqw 0:2847; Kdw 2:6858; Kpdc 0:1805; Kqdc 0:6112;
Kvm 0:8882; Kvw 0:0138; Kvdc 0:2994
166

DIdc2
1
DVdc
rbess
4:125
From Eqs. (4.19) and (4.125), state equation of the example power system is
obtained as
X_ AX Bu
4:126
where
2
Dd
7
6
6 Dx 7

7
6
D/
0 7
6
;
X 6 DEq 7; u
Dm
7
6
4 DE0fd 5
DVdc
3
2
0
xo
0
0
0
7
6
K
D
1
2
K
M
0
Mpdc 7
6 K
M
M
7
6
7
6
Kqdc
K3
1
7
6 K0 4
0

0
0
A6
7
T
T
T
T
d0
d0
d0
d0
7
6
7
6 KA K5
K
K
1
A K6
vdc
7
6
0
T
T
T
TA
5
4
A
A
A
K7
0
K8
0
K9 C 1r
dc bess
2
0
314:16
0
0
0
6
0
0:0838
0
0:0226
6 0:1302
6
6
0
0:3965 0:1983
0:1212
6 0:0930
6
0
5483:2 100
2994:2
4 428:242
0:7086
2
0
6 Kpw
6
M
6
6
Kqw
0
B6
Td0
6
6
6 KA Kvw
4 T
A
Kdw
0:4067
0
100:0746
3
0
3
2
0
0
Kpm 7
M 7 6
7
7 6 0:0886 0:0669 7
7
6
7
K
7
7 6
qm
T0d0 7 6 0:0564 0:3594 7
7
7 6
8882 5
7 4 138:495
KATKvm 5
A
2:6858 0:2212
Kdm
0
3
7
7
7
7;
7
7
5
4.4 Examples
167
Eigenvalues of state matrix are

k1 ; k2 0:0591 j6:4978
k3 13:0310
k4 100:2902
k5 87:2679
k1 and k2 are the pair of electromechanical oscillation mode of the example power
system.
4.4.1.2
Design of BESS Stabilizers by Use of the Phase

Compensation Method
To work out the transfer function of the forward path from a BESS stabilizer to the
electromechanical oscillation loop, the path related to Dd in the HeffronPhillips
model of Figs. 4.4 and 4.5 can be ignored. Hence, from Fig. 4.4 with Dd being
ignored, it can have
DVdc K8 DE0q Kdm Dm Kdw D/
1
s K9 C 1r
dc bess
4:127
Figure 4.19 shows the forward path from the BESS to the electromechanical
oscillation loop of generator. From Fig. 4.19, it can be obtained

1
KA
Kvdc DVdc Kvw Dw Kvm Dm K6 DE0q
sT0d0 K3 1 sTA

Kqdc DVdc Kqw D/ Kqm Dm
DE0q
4:128
DTet K2 DE0q Kpdc DVdc Kpw D/ Kpm Dm
Tet
[ Vdc
m ]
K2
K pdc
K p
K pm
K qdc
E q '
1
K 3 + sTd0 '
K qm
K v
K vm
K q
K6
Fig. 4.19 Forward path of BESS
K vdc
KA
1 + sTA
168
Deleting variable DVdc and DE0q in Eqs. (4.126) and (4.127), it can be obtained
that
DTes Fact sD/ Frct sDm
4:129
With xs Imk1 or k2 6:4978, it can be computed that

Fact jxs 0:7164 j0:0017 0:7164\ 0:136
Frct jxs 0:2467 j0:4706 0:5313\117:661
4:130
A simple way to calculate the transfer function of the forward path with s jxs
is to obtain the following state-space representation of the system of Eqs. (4.127)
and (4.128)
X_ r Ar Xr Br u
DTet Cr Xr Dr u
4:131
3

DE0q
D/
, Ar and Br are obtained directly from A and
where Xr 4 DE0fd 5; u
Dm
DVdc
B from Eq. (4.126) by deleting two state variables Dd and Dx, and the output
equation is obtained from the second equation in Eq. (4.128). That is
3
Kqdc
3
2
0:3965 0:1983
0:1212
Td0 0
Td0 0
d0
7
6
7
7 6
6
Ar 6 KA K6 1
4 5483:2 100
2994:2 5;
KTvdc 7
TA
TA
5
4
A
0:4067
0
100:0746
K8
0
K9 C 1r
dc bess
3
2
3
2
K
K
T pw0
T pm0
0:0564 0:3594
7
6
d0
d0
7
7 6
6
8882 5;
Br 6 KA Kvw KA Kvm 7 4 138:495
5
4
TA
TA
2:6858 0:2212
Kdw
Kdm
3
3 2
2
K2
0:6702
7
7 6
6
Cr 4 0 5 4 0 5; Dr Kpw Kpm 0:7086 0:5355
Kpdc
0:1805
2
TK30
It can have
DTet CTr sI Ar 1 Br Dr u Fact sD/ Frct sDm
4:132
4.4 Examples
169
Hence,
Frct jxs CTr jxs I Ar 1 Br Dr
0:7164\ 0:136 0:5313\117:661
Fact jxs
The above results are as same as those given in Eq. (4.130). Take the transfer
function of the BESS stabilizer to be
Tbess s Kbess
1 sT2
1 sT1
with T1 0:38s:
Deviation of active power delivered along the transmission line, DPt , is taken as the
feedback signal of the stabilizer. Hence, the electric torque supplied by a BESS
active power stabilizer is
DTetactive Fact jxs D/ Fact jxs Tbess jxs DPt
Fact jxs Tbess jxs D jxs MDx
4:133
The electric torque supplied by a BESS reactive power stabilizer is

DTetreactive Frct jxs Dm Frct jxs Tbess jxs DPt
Frct jxs Tbess jxs D jxs MDx
4:134
The BESS active and reactive power stabilizer can be designed by using the phase
compensation method to supply a pure damping torque DTetactive 9Dx and
DTetreactive 20Dx, respectively. The parameters of stabilizers obtained from the
design are as follows:
1. The BESS active power stabilizers: Kbess 15:9567; T2bess 0:3826s:;
2. The BESS reactive power stabilizers: Kbess 99:7980; T2bess 0:0151s:
Table 4.1 gives the computational results of oscillation mode of example power
system without and with the BESS stabilizers installed. They conrm the effectiveness of the BESS stabilizer in damping the power oscillations.
Figure 4.20 shows the simulation results of the example power system without
and with the BESS stabilizers installed. In the simulation, a three-phase to-earth
Table 4.1 Oscillation mode of example power system when Pt0 = 0.5 p.u.
Without any stabilizer
With active power stabilizer
With reactive power stabilizer
0:0591 j6:4978
0:6541 j6:4621
0:5315 j5:1291
170
Rotor angle (degree)
with reactive stabilizer installed
with active stabilizer installed

without stabilizers
time(second)
Fig. 4.20 Simulation results of example power system when Pt0 0:5 p.u.
short circuit occurred at 0.5 s of simulation for 100 ms. The results conrm the
effectiveness of the BESS stabilizers designed by use of the phase compensation
method to damp the power oscillation.
4.4.1.3
Robustness of the BESS Stabilizers to the Variations of Power

System Loading Conditions
Effectiveness of the BESS stabilizers is examined when Pt0 varies from 0.1 p.u. to
1.0 p.u. with Vt0 1:0; Vb 1:0; Vs0 1:0 being xed. Figure 4.21 shows the
damping torque provided by the BESS stabilizers with variations of system load
conditions. From Fig. 4.21, it can be seen that provision of the damping torque by
the BESS active power stabilizer changes much less than that by the reactive power
stabilizer. Hence, the active power stabilizer is more robust to the variations of
system load conditions. Table 4.2 gives the computational results of electromechanical oscillation mode. Figures 4.22 and 4.23 are the results of simulation of
example power system at different load conditions. Those results obviously conrm
that the BESS active power stabilizer is more robust than the reactive power stabilizer to the variations of power system loading conditions as concluded by the
discussion in Sect. 4.1.2.2.
4.4 Examples
171
Damping torque supplied by stabilizer

Operating condition
Fig. 4.21 Damping torque provided by BESS stabilizers with variations of system loading
conditions
Table 4.2 Oscillation mode of example power system with variable loading conditions
With active power stabilizer
With reactive power stabilizer
Pt0 = 0.2 p.u.
Pt0 = 0.9 p.u.
0:0070 j6:3064
0:6685 j6:2676
0:2050 j5:7870
0:2505 j6:8670
0:5427 j6:8615
0:9353 j4:6027

without stabilizers
time(second)
172

without stabilizers
time(second)
4.4.2
An Example Single-Machine Innite-Bus Power

System Installed with a UPFC Stabilizer
4.4.2.1
Extended HeffronPhillips Model of Example Power

System Installed with a UPFC
Conguration of an example single-machine innite-bus power system installed

with a UPFC is as same as that shown in Fig. 4.12. Parameters of example system
are as follows:
Generator
The AVR
Transmission
line
The UPFC
Steady-state
operating point
xd 1:0; xq 0:6; x0d 0:3; M 8s:; D 0; T0d0 5:044s:

KA 100; TA 0:01s:
xtb 0:3; xte 0:15; xbt 0:15:
ke 3; kb 3; Cdc 1:0; Vdc0 1:0; xes 0:02; xbs 0:02:
Pt0 0:1; Vt0 1:05; Vb 1:0:
This example is used to demonstrate the design and effectiveness of the UPFC
stabilizers. For the simplicity of presentation, normal control functions of the UPFC
are not included in the following demonstration. The procedure will be exactly as
4.4 Examples
173
same as what is presented below when the UPFC normal control functions are
considered. Hence, it is assumed that Ies 0; Vbt 0 which can give
Vbs jxbs Ibs ;
Ves Vet
4:135
Total reactance of the transmission lines is xt xxtbtb xxte texxbtbt 0:15. From

Vb
Vt0 Vb
jxt
Pt0 jQb0
, the reactive power received at the innite bus bar Qb0 can be
obtained as
Qb0
2s
3

2
Vb 4
xt Pt0
V2t0
Vb 5 0:3326
xt
Vb
4:136
Thus,
Pt0 jQb0
0:1 j0:3326
Vb
Vt0 jxt Itb0 Ite0 Vb 1:0499 j0:0150 1:05\0:82
xtb
Ite0 Ibs0 Itb0 Ite0
0:05 j0:1663 0:1737\ 73:27
xtb xte xbt
EQ Vt0 jxq Itb0 Ite0 1:2495 j0:0750 1:252\3:435
Itb0 Ite0
Ves0 Vet0 Vt0 jxte Ite0 1:0249 j0:0075 1:025\0:419 me0 ke Vdc0 \86:99
Vbs0 jxbs Ibs 0:0033 j0:001 0:0034\16:86 mb0 kb Vdc0 \86:56
Hence, d0 3:435 ; me0 0:3417; mb0 0:0012; de0 0; db0 2:8459 ,

vtd0 ReVt0 sin d0 ImVt0 cos d0 0:0479;
vtq0 ReVt0 cos d0 ImVt0 sin d0 1:0489;
ited0 ReIte0 sin d0 ImIte0 cos d0 0:1690;
iteq0 ReIte0 cos d0 ImIte0 sin d0 0:0399;
vtq0 Vb cos d0
itbd0
0:1690;
xtb
Vb sin d0 vtd0
0:0399;
itbq0
xtb
vetd0 ReVet0 sin d0 ImVet0 cos d0 0:0539;
vetq0 ReVet0 cos d0 ImVet0 sin d0 1:0236;
iesd0 0; iesq0 0;
ibsd0 ReIbs0 sin d0 ImIbs0 cos d0 0:1690;
ibsq0 ReIbs0 cos d0 ImIbs0 sin d0 0:0399;
E0q0 EQ xd x0d ited0 itbd0 1:1503
174
Various coefcient matrices in Eqs. (4.86) and (4.88) are computed to be

Fq
0:9805
0:4001

;
0:0367
1:4073
Fd
0:0134 0 0:2689
0:0408 0:1656

1:1742
7:6144 0:9001 0:0174
0:7903 5:1251 0:9515

0:0184

23:3120 0:4196
14:8739 0:001
14:8800 0:2678 15:8655 0:0011

0:8589 0 0:0262
0:0768 0:4979 0:0103 0:0002
Gq
Gd
Hd
Ld
Hq 0:0416
Lq 1:2648
7:9475
5:0655
0:4323
0:2641
0:0228
0
0:1487
Following state equation of extended HeffronPhillips model of Eq. (4.95) is

obtained
2
Dd_
Dx_
0
DE_
314:1593
6
7 6
6
7 6 0:2244
0
6
7 6
6
7 6
0
6
q 7 6 0:0192
6
7
6 _0 7 6
23:2301
0
4 DEfd 5 4
1:0061
0
DV_ dc
2
0
0
6
6 0:0074 0:1301
6
6 0:5851 0:0105
6
6
0:1463
4 12670
0:0549
7:7987
0:0173
32
Dd
7
76
0:0025 76 Dx 7
7
76
0 7
6
0:5387 0:1983 0:2000 7
76 DEq 7
7
76
2638:0 100 4330:5 54 DE0fd 5
DVdc
0:0759
0
0:0189
3
0
0
2
3
7 DmE
0:0012 0:00005 76
76 DdE 7
7
4:137
0:0688 0:000004 7
7
76
4
Dm
B5
7
1490:5
0:0127 5
DdB
0:0308
0:0172
Eigenvalues of state matrix are

k1 94:4375
k2 ; k3 0:0106 j8:4249
k4 6:0714
k5 0:0276
Obviously, k2 ; k3 0:0106 j8:4249 are the pair of electromechanical
oscillation mode of example power system.
4.4 Examples
4.4.2.2
175
Robust Selection of a Modulation Signal to Add the Damping

Control Signal of the UPFC Stabilizer by Using the Damping
Torque Analysis and Residue Index
With Vt0 1:05; Vb 1:0, Pt0 changes from 0.1 p.u. to 1.2 p.u. Deviation of
modulation signal to add a UPFC stabilizer is Duk ; k 1; 2; 3; 4 where
Du1 Dme ;
Du2 Dde ;
Du3 Dmb ;
Du4 Ddb
4:138

According to the denition of Eq. (4.108) and from Fig. 4.17, Fupfck jxs ; k
1; 2; 3; 4 is calculated as shown in Fig. 4.24 (jxs j8:4249).
From Fig. 4.24, it can be seen that

rselected fPt0 0:1g MinFupfck jxs ; k 1; 2; 3; 4; r 2 fPt0 0:1 1:2g
r
It means that the UPFC stabilizer should be designed at load condition Pt0 0:1
when the stabilizer is least effective to damp the electromechanical oscillation
mode. It is also shown in Fig. 4.23 that the stabilizer can provide the most amount
of damping torque when it is added on the modulation signal Dde . That is

Duselected Dde MaxFupfck jxs ; k 1; 2; 3; 4; Duk 2 fDme ; Dmb ; Dde ; Ddb g
Duk
Magnitude of Forward path
with e signal
with m e signal
with m b signal
with b signal
Operating condition

Fig. 4.24 Computational result of Fupfck jxs ; k 1; 2; 3; 4
176
Hence, to have a robust UPFC stabilizer, the stabilizer should be added on Dde
and the design carried out at operating condition Pt0 0:1; Vt0 1:05; Vb 1:0.
The robust modulation signal to add a UPFC stabilizer can also be selected by
using the residue index dened by Eq. (4.104). Take the feedback signal of stabilizer to be the locally available signal, the deviation of line active power Pte .
Similar to Eq. (4.83), it has
Pte vtq iteq vtd ited
4:139
By using Eqs. (4.86) and (4.88), similar to the derivation of Eq. (4.91), linearization of Eq. (4.139) can be obtained to be
DPte iteq0 Dvtq ited0 Dvtd vtq0 Diesq Dibsq vtd0 Diesd Dibsd
3
2
Dd
3
2
Dme
7
6
6 Dx 7
6 Dd 7
7
6
6 e7
0 7
c1 0 c3 0 c5 6
7
6 DEq 7 d1 d2 d3 d4 6
4 Dmb 5
7
6
4 DE0fd 5
Ddb
DVdc
4:140
where
c1
d1
c3
d2
0
d3
c5 Itq0 Hq Itd0 Hd Vtq0 Fq1 Vtq0 Fq2

Vtd0 Fd1 Vtd0 Fd2 1:1616 0 0:0859 0
d4 Itq0 Lq Itd0 Ld Vtq0 Gq1 Vtq0 Gq2
Vtd0 Gd1 Vtd0 Gd2 0:0620 2:7016
0:0212
0:0141
0:0011
Fq1 and Fq2 , Fd1 and Fd2 , Gq1 and Gq2 and Gd1 and Gd2 are the rst and
second rows of Fq , Fd , Gq , and Gd , respectively.
At the selected operating condition rselected fPt0 0:1; Vt0 1:05;
Vb0 1:0g, for the oscillation mode k2;3 0:0106 j8:4249, the right and left
eigenvectors are obtained as
2
0:0091 j0:0206
0:0091 j0:0206
6
6
7
7
6 0:7702 j0:3380 7
6 0:7702 j0:3380 7
6
6
7
7
6
7
7
v2 6
6 0:0013 j0:0008 7 and v3 6 0:0013 j0:0008 7
6
6
7
7
4 0:0000 j0:0000 5
4 0:0000 j0:0000 5
0:0010 j0:0010
0:0010 j0:0010
3
3
2
2
0:0899 j0:2015
0:0899 j0:2015
7
7
6
6
6 0:0054 j0:0024 7
6 0:0054 j0:0024 7
7
7
6
6
7
7
3 6
2 6
w
6 0:0021 j0:0224 7 and w
6 0:0021 j0:0224 7
7
7
6
6
0:9747
0:9747
5
5
4
4
0:0243 j0:0107
0:0243 j0:0107
4.4 Examples
177
From Eqs. (4.102) and (4.103), the controllability of four input signals as far as
the electromechanical oscillation mode is concerned is calculated as
k 1; Du1 DmE ;

T2 b1 2:4108 j2:3637;
b21 w
k 2;

b22

b23
Du2 DdE ;
T2 b2
w
k 3;
b31 w
T3 b1 2:4108 j2:3637
10:7641 j5:1591;
b32 w
T3 b2 10:7641 j5:1591
Du3 Dmb ;
T2 b3
w
b33 w
T3 b3 0:2548 j0:2928
0:2548 j0:2928;
k 4; Du4 Ddb ;

T2 b4 0:0023 j0:0001;
b24 w
b34 w
T3 b4 0:0023 j0:0001
where
2
2
2
2
3
3
3
3
0
0
0
0
6 0:0074 7
6 0:1301 7
6 0:0012 7
6 0:0000 7
6
6
6
6
7
7
7
7
6
6
6
7
7
7
7
b1 6 0:5851 7; b2 6 0:0105 7; b3 6 0:0688 7; b4 6
6 0:0000 7;
4 12670 5
4 0:1463 5
4 1490:5 5
4 0:0127 5
0:0549
7:7986
0:0308
0:0172
The observability is given by
c2k cTk
v2 0:0002 j0:0029
and c3k cTk v3 0:0002 j0:0029
where
cTk 1:1616
0:1387
0 0:0202
k 1; 2; 3; 4
The residue index for four input signals fDmE ; DdE ; DmB ; DdB g of the UPFC
stabilizer, thus, is obtained as
k 1; Du1 DmE ;
R21
b21c21 0:0073 j0:0065;
R31 b31c31 0:0073 j0:0065;
k 2; Du2 DdE ;
R22
b22c22 0:0128 j0:0322;
R32 b32c32 0:0128 j0:0322;
k 3; Du3 Dmb ;
R23
b23c23 0:0009 j0:0007; R33 b33c33 0:0009 j0:0007;
k 4; Du4 Ddb ;
24c24 0:0000 j0:0000; R34 b34c34 0:0000 j0:0000
R24 b
Because R2k and R3k (k = 1, 2, 3, 4) are pairs of conjugate vectors, and the
amplitude of residue measures the overall influence of the UPFC stabilizer on the
oscillation mode. It can be obtained that
178
Residue
with
m e signal
with e signal
with
m b signal
with b signal
Pt0
Fig. 4.25 Computational result of jRki j; k 1; 2; 3; 4

R21 R31 0:0097

R22 R32 0:035

R23 R33 0:0011

R24 R34 0
The largest residue indicates the most effective feedback signal to design the UPFC
stabilizer. Hence,

uselected dE where Rik max Ri2 0:035i 2; 3
The selection is as same as that obtained previously by using the damping torque
analysis. Variation of the residue with system load condition from 0.1 to 1.2 p.u. is
shown in Fig. 4.25. Obviously, Pt0 0:1 should be selected, which again is as same
as that obtained previously.
4.4.2.3
Design of the UPFC Stabilizer by Using the Phase

Compensation Method
State Eq. (4.137) and output Eq. (4.140) are obtained at the selected operating
condition Pt0 0:1; Vt0 1:05; Vb 1:0 of the example power system.
With Dde being selected to add the damping control signal of the UPFC stabilizer, it can be obtained that
4.4 Examples
179
k 0:2244; d 0; C1 1:1616; C2 0; C3 0:0859 0

dm 2:7016
A23 0:0173 0 0:0025
3
2
2 3
0:0192
0
7
6
6 7
A31 4 23:2301 5; A32 4 0 5;
1:0061
A33
0:5387
6
4 2638:0
0:1983
100
0:0759
0:0212 ;
3
0:2
7
4330:5 5
0:0189
B2 0:1301
2
3
0:0105
6
7
B3 4 0:1463 5
7:7986
Results of computation according to Eq. (4.61) are (jxs j8:4249)

jxs
jxs
Kjxs k
d AT23 jxs I A33 1 A31
A32 0:2259 j0:0006
x0
x0

x0
x
0
Ko jxs
C1 C2 CT3 jxs I A33 1
A31 A32 0:0663 j43:5996
jxs
jxs
Kc jxs AT23 jxs I A33 1 B3 B2 0:1418 j0:0047
Kil jxs CT3 jxs I A33 1 B3 dm 2:7599 j0:0155
Let the design of the UPFC stabilizer provide a pure damping torque
Dupfc Dx 1Dx. Take DPte as the feedback signal. Thus,
Dde Tupfc sDPte Kupfc
1 sT2
DPte
1 sT1
4:141
where T1 0:3 s.
From Eq. (4.64), the forward path of the stabilizer is calculated as
Fupfc jxs Kc jxs Ko jxs Dupfc Kil jxs 2:9764 j6:1674
Thus, parameters of the UPFC stabilizer are obtained as
Kupfc 0:2689;
T2 0:1288:
Table 4.3 gives the computational results of the electromechanical oscillation

mode of example power system at three different load conditions. Figures 4.26,
4.27, and 4.28 are the results of nonlinear simulation. Those results conrm that the
UPFC stabilizer is effectively designed in damping the power oscillation and robust
to the variations of system load conditions.
180
Table 4.3 Oscillation mode of example power system with variable load conditions
Without stabilizer
With stabilizer
Pt0 0:1 p.u.
Pt0 0:8 p.u.
Pt0 1:2 p.u.
0:0106 j8:4249
0:6880 j7:0257
0:0133 j8:9982
1:2283 j7:5835
0:1457 j9:4250
1:2930 j8:1842
(degree)
without stabilizer
with UPFC stabilizer
time(second)
Fig. 4.26 Simulation results of example power system when Pt0 = 0.1 p.u.
(degree)
without stabilizer
time(second)
4.4 Examples
181
(degree)
without stabilizer
time(second)
In order to demonstrate that the robustness of the UPFC stabilizer is due to the
selection of the load condition at Pt0 0:1; Vt0 1:05; Vb 1:0 where the
stabilizer is designed, the stabilizer is designed at another load condition
Pt0 0:8; Vt0 1:05; Vb 1:0. Then, its effectiveness is checked at different
load conditions by modal computation presented in Table 4.4. From the results, it
can be seen that though the UPFC stabilizer is designed properly at the selected
load condition Pt0 0:8; Vt0 1:05; Vb 1:0, where the power oscillation is
effectively suppressed (see the second column of Table 4.4), its effectiveness is
not ensured at Pt0 0:1; Vt0 1:05; Vb 1:0 (see the rst column of
Table 4.4).
At the selected operating condition Pt0 0:1; Vt0 1:05; Vb 1:0, the UPFC
stabilizer can also be added to other modulation signals. Table 4.5 gives the results
of designing the UPFC stabilizers added on other two modulation signals and the
oscillation mode of example power system when stabilizers are installed. From
Table 4.5, it can be seen that the gain value of the UPFC stabilizers added on Dme
and Dmb is much greater than that of the stabilizer added on Dde in order to
Table 4.4 Oscillation mode of example power system with the stabilizer designed at
Pt0 0:8; Vt0 1:05; Vb 1:0
Without stabilizer
With stabilizer
Pt0 0:1 p.u.
Pt0 0:8 p.u.
Pt0 1:2 p.u.
0:0106 j8:4249
0:4291 j6:8578
0:0133 j8:9982
0:7517 j7:3203
0:0133 j8:9982
0:7565 j7:5281
182
Table 4.5 Results of designing the UPFC stabilizers added on other two modulation signals
Modulation signal added with
a UPFC stabilizer
Parameters of the UPFC

stabilizer
Oscillation mode with the

Dme
Dmb
Kp 1:6654; Tp 0:0926
Kp 15:3864; Tp 0:1069
0:7030 j6:9400
0:7078 j7:0774
suppress the power oscillation effectively. This means that more control cost is
required. Hence, the UPFC stabilizer added on Dde is the most efcient damping
controller.
References
1. CIGRE TF 30-01-08 Report, Modelling of power electronics equipment (FACTS) in load flow
and stability programs (1999)
2. Wang HF (1999) Phillips-Heffron model of power systems installed with STATCOM and
applications. IEE Proc Part C 146(5):521527
3. Du W (2009) Power system small signal oscillation stability as affected by static synchronous
compensator (STATCOM) and energy storage system (ESS). PhD thesis, University of Bath,
UK
4. Wang HF (2000) Static synchronous series compensation to damp power system oscillations.
Int J Power Syst Res 54(2)
5. Wang HF (1999) Applications of modelling UPFC into multi-machine power systems. IEE Proc
Part C (3)
6. Wang HF (1999) Damping function of unied power flow controller. IEE Proc Part C. (1)
7. Wang HF (2000) A unied model for the analysis of FACTS devices in damping power system
oscillations part III: unied power flow controller. IEEE Trans Power Delivery (3)
Chapter 5
A Multi-machine Power System Installed

with Power System Stabilizers
5.1
Mathematical Model of a Multi-machine Power

System Installed with Power System Stabilizers
5.1.1
A Two-Machine Power System Installed with Power

System Stabilizers
Figure 5.1 shows the conguration of a two-machine power system where two
synchronous generators are connected by a transmission line. This simple system is
used in this section to demonstrate the procedure of establishing the mathematical
model of a multi-machine power system. Later, it will be used for introducing basic
concepts and methods in the analysis of power system oscillations and design of
PSSs in the multi-machine power system. For the purpose of demonstration, a load is
connected at the terminal of two generators and a node along the transmission line,
respectively. The load is represented by equivalent impedance rLi jxLi ; i 1; 2; 3:
5.1.1.1
Network Equations
Connections of two synchronous generators in the two-machine power system are

given by following network equations expressed in a common xy coordinate
1
g1
V
rL1 jxL1
1
Ig2
g2
V
rL2 jxL2
Ig1
1
3
3 gL1 jbL1 jb13 V
g1 jb13 V
Vg1 V
jx13
5:1
1
3
3 gL2 jbL2 jb23 V
g2 jb23 V
Vg2 V
jx23

DOI 10.1007/978-1-4899-7696-3_5
183
5 A Multi-machine Power System Installed
184
G1
Vg1
Ig1
V3
x13
x 23
Vg 2
rL3 + jx L3
rL1 + jx L1
Ig 2
G2
rL2 + jx L2
Fig. 5.1 A two-machine power system
1
3 1 V
3 V
g1 1 V
3 V
g2
V
rL3 jxL3
jx13
jx23
g1 jb23 V
g2
3 jb13 V
gL3 jbL3 jb13 jb23 V
5:2
where jb13 jx1 ; jb23 jx1 ; gLi jbLi r 1jx ; i 1; 2; 3:

Li
13
23
Li
Matrix form of above equations is
2 3 2
gL1 jbL1 jb13
Ig1
4 Ig2 5 4
0
jb13
0
0
gL2 jbL2 jb23
jb23
32
jb13
56
jb23
4
gL3 jbL3 jb13 jb23
g1 3
V
g2 7
V
5
3
V
5:3

3
V
jb13
jb23
g1
g2
V
V
gL3 jbL3 jb13 jb23
gL3 jbL3 jb13 jb23
5:4
Substituting above equation into Eq. (5.1), it can be obtained that

"
Ig1
Ig2
3
jb jb
jb13 jb23
"
#
gL1 jbL1 jb13 g 13 jb13

gL3 jbR
L3
R
7 Vg1
6
4
5
g2
jb23 jb13
jb23 jb23
V
g
jb
jb
L2
L2
23
gL3 jbR
gL3 jbR
"
#"
#
"
#
g1
g1

yN11 yN12
V
V
N
Y
g2
g2

yN21 yN22
V
V
2
5:5
5.1 Mathematical Model of a Multi-machine

Fig. 5.2 Phasor diagram in
xy coordinate and di qi
coordinate
185
qi
y
E qi '
(x qi x di ')Iqi
jx di ' Igi
V gi
di
N is the network admittance matrix with only

where bLR bL3 b13 b23 and Y
nodes of generator terminal left. Equation (5.2) can be rearranged as
3 2
igx1
gN11
6 igy1 7 6 bN11
7 6
6
4 igx2 5 4 gN21
igy2
bN21
2
bN11
gN11
bN21
gN21
gN12
bN12
gN22
bN22
32
3
vgx1
bN12
7
6
gN12 7
76 vgy1 7
5
4
vgx2 5
bN22
vgy2
gN22
5:6
where subscript x and y denotes x and y component of a variable in common xy

coordinate, respectively. According to the phasor diagram of Fig. 2.9, in dq
coordinate of the ith generator, di -qi , it can have (i 1; 2)
gi E
0 jx0 Igi xqi x0 iqi
V
qi
di
di
5:7
0 is
From Fig. 5.2, it can be seen that in the common xy coordinate, phase of E
qi
0

di and that of xiq xid iqi is di 90 . Hence, in xy coordinate, Eq. (5.7) is
(i 1; 2)
gi E0 ejdi jx0 Igi xqi x0 iqi ejdi 90
V
qi
di
di
5:8
186
In matrix form, Eq. (5.8) is

"
g1
V
g2
V
"
#"
#
" #
E0q1
Ig1
0
jx0d1
ejd1
0

0
0
jd2
Eq2
0
jxd2 Ig2
0 e
"
#

iq1
xq1 x0d1 iq1 ejd1 90
0
0
jd2 90
iq2
0
xq2 xd2 iq2 e
5:9
Denote
gi E0 ejdi xqi x0 iqi ejdi 90 ;
E
qi
di
i 1; 2
5:10
as an internal voltage of generator. Equation (5.9) becomes

0
g1
g1
E
V
jxd1
g2
g2 0
E
V
0
jx0d2
Ig1
Ig2

5:11
gi ; i 1; 2 can be
Equation (5.11) indicates that introduction of internal voltage E
seen as addition of an extra node behind the node of generator terminal as shown in
Fig. 5.3. The added extra internal voltage node is connected to the generator
terminal through x0di ; i 1; 2.
Ig1
Ig2

0
jxd1
1

YN
0
0
jx0d2
1
g1
E
g2
E

g1
E

Y
Eg2
5:12
is the network admittance matrix with only internal voltage nodes

where matrix Y
of generators left.
E g1
x d1 '
Vg1
Ig1
rL1 + jx L1
x13
V3
x 23
rL3 + jx L3
Fig. 5.3 Introduction of internal voltage nodes of generators
Ig 2
Vg 2
Eg 2
xd2 '
rL2 + jx L2
187
That is i 1; 2
Igi igxi jigyi
2
X
gj
yij E
j1
2
X
yij E0qj ejdj xqj x0dj iqj ejdj 90
5:13
j1
According to Fig. 5.2, in di -qi coordinate

where
yij yij ejaij is the element of Y.
Eq. (5.13) is (i 1; 2)
Igi idi jiqi ixi jiyi ej90
di
2
X
yij E0qj ejdj xqj x0dj iqj ejdj 90 ej90
di
j1
2
X
dj di
dj di aij

yij E0qj ej90
xqj x0dj iqj ejdj di
j1
2
X
yij E0qj ej90
xqj x0dj iqj ejdj di aij
j1
2
X
yij E0qj cos(90 dj di aij jE0qj sin(90 dj di aij
j1
xqj x0dj iqj cos(dj di aij j(xqj x0dj iqj sin(dj di aij
2
X
yij E0qj sin(dj di aij jE0qj cos(dj di aij
j1
xqj x0dj iqj cos(dj di aij j(xqj x0dj iqj sin(dj di aij
5:14
From above equation, it can have i 1; 2
idi
2
X
yij E0qj sindj di aij xqj x0dj iqj cosdj di aij
j1
iqi
2
X
5:15
yij E0qj cos(dj
di aij xqj
x0dj iqj sin(dj
di aij
j1
5.1.1.2
Linearized Model When the Classical Model of Synchronous

Generators Is Used
For the simplicity of discussion, rstly the simplest case is considered when a
synchronous generator is represented by the so-called classic model, i.e., the rotor
motion equation of generator. That is the model given by Eq. (2.41) with function
of the AVR ignored and E0q being constant. In di -qi coordinate, model of generator
is (i 1; 2)
188
d_ i xo xi 1
1
x_ i
Pmi Pi Di xi 1
Mi
Pi vdi idi vqi iqi
5:16
vdi xqi iqi

vqi E0qi x0di idi
Hence, linearized classical model of generator is (i 1; 2)
Dd_ i xo Dxi
1
Dx_ i
DPi Di Dxi
Mi
DPi vdi0 iqi0 x0di Didi vqi0 idi0 xqi Diqi
5:17
Linearization of Eq. (5.15) is (E0qi is a constant, i; j 1; 2)

Didi hdi Ddj Ddi
Diqi hqi Ddj Ddi
5:18
i 6 j
where hdi and hdi i 1; 2 are constants. Hence, from Eqs. (5.17) and (5.18), it can
have (i; j 1; 2)
DPi ki Ddj Ddi
5:19
where ki i = 1, 2 are constants. State equation of the two-machine power system is

3 2 0
Dd_ 1
0
6 _ 7 6
6 Dd2 7 6
k
1
6
76
4 Dx_ 1 5 6
4 M1
k2
Dx_ 2
M
2
2
0
0
k1
M
1
k2
M2
xo
0
D1
M
1
0
32
2
3
3
0
Dd1
Dd1
xo 7
6 Dd 7
76
Dd2 7
6 27
7
76
A
6
7
7
46
0 74 Dx 5
4 Dx1 5
1
5
D2
Dx2
Dx2
M
5:20
Equation (5.20) is equivalent to the following equation

0
D1
@s2 I sxo 4 M1
0
3
2
k1
0
5 xo 4 M1
D2
k2
M
M2
2
31

k1
M
Dd1
1 5A
0
k2
Dd2
M2
5:21
Characteristic equation of the system is

2
D 1 x k1
k1
xo M

k kxo M
oM
1
1
1
0

2
k
D
k
xo M2
k kxo M2 xo M2

2
2
2
189
5:22
Obviously, k 0 is one of the eigenvalues of the system. Because state matrix

A4 is a real matrix, if there is a power oscillation, other three eigenvalues must be a
pair of conjugate complex eigenvalues and a real eigenvalue. The pair of conjugate
complex eigenvalues is the electromechanical oscillation mode, because the only
dynamic of system is the rotor motion of generators. Therefore, in the two-machine
power system, there is only one electromechanical oscillation mode. In fact, if it is
denoted that Dd12 Dd1 Dd2 , the state equation of system of Eq. (5.20) becomes
3 2 0
Dd_ 12
6
7 6 k1
4 Dx_ 1 5 6
4 M1
k2
Dx_ 2
M2
2
xo
D1
M
1
0
3
2
3
2
3
Dd12
Dd12
7
0 74 Dx 5 A 4 Dx 5
1
3
1
5
D2
Dx2
Dx2
M
2
0
5:23
Obviously, the system can only have one pair of conjugate complex eigenvalues
and hence one electromechanical oscillation mode. The reduced-order state matrix
A3 does not have to have one zero eigenvalue when the relative angular position of
generators is used as the state variable.
From Eq. (5.19), it can be seen that variations of the active power delivered
along the transmission line connecting two generators are proportional to the relative angular position. For example, if Dd1 Dd2 , DPi 0 i 1; 2, there will be
no variations of active power even if the absolute angular positions of generators
vary. This means that the power oscillations are related only to the relative angular
positions of generators.
5.1.1.3
HeffronPhillips model of the two-machine power system installed with the PSS
can be established based on the following 4th-order model of synchronous generators i 1; 2
d_ i xo xi 1
1
x_ i
Pmi Pi Di xi 1
Mi
1
0
E_ qi 0 Eqi Efd0i E0fdi
Tdoi
1 0
KAi
0
E_ fdi
E
Vrefgi Vgi upssi
TAi fdi TAi
5:24
190
where
vid xqi iqi
vqi E0qi x0di idi
Pi vdi idi vqi iqi
5:25
Eqi E0qi xdi x0di idi

q
Vgi v2di v2qi
Linearization of Eq. (5.15) is (i 1; 2)
Didi
2
X
hdij DE0qj hdj Ddj Ddi
j1
Diqi
2
X
hqij DE0qj hqj Ddj Ddi
5:26
j1
i 6 j
By using Eq. (5.26), linearization of Eq. (5.25) can be obtained as
DP1 k11 Dd2 Dd1 k211 DE0q1 k212 DE0q2
DP2 k12 Dd1 Dd2 k221 DE0q1 k222 DE0q2
DEq1 k41 Dd2 Dd1 k311 DE00q1 k312 DE0q2
DEq2 k42 Dd1 Dd2 k321 DE0q1 k322 DE0q2
5:27
DVg1 k51 Dd2 Dd1 k611 DE00q1 k612 DE0q2

DVg2 k52 Dd1 Dd2 k621 DE00q1 k622 DE0q2
By substituting Eq. (5.27) into the linearization of Eq. (5.24), it can have for G1
Dd_ 1 xo Dx1
1
k11 Dd2 Dd1 k211 DE0q1 k212 DE0q2 D1 Dx1
M1
1
0 k41 Dd2 Dd1 k311 DE0q1 k312 DE0q2 DE0fd1
Tdo1
Dx_ 1
0
DE_ q1
0
DE_ fd1
1
KA1
DE0fd1
k51 Dd2 Dd1 k611 DE0q1 k612 DE0q2 Dupss1
TA1
TA1
5:28
191
and for G2
Dd_ 2 xo Dx2
1
Dx_ 2
k12 Dd1 Dd2 k221 DE0q1 k222 DE0q2 D2 Dx2
M2
1
0
DE_ q2 0 k42 Dd1 Dd2 k321 DE0q1 k322 DE0q2 DE0fd2
Tdo2
1
KA2
0
DE_ fd2
DE0
k52 Dd1 Dd2 k621 DE0q1 k622 DE0q2 Dupss2
TA2 fd2 TA2
5:29
Equations (5.28) and (5.29) are the HeffronPhillips model of the two-machine
power system installed with PSS as shown in Fig. 5.4.
State-space representation of the HeffronPhillips model is
2
2
0
7 6
7 6 M1 K1
74
T1
q 5
d0 K4
1
0
T
K5 KA
_
A
DEfd
:
Dd
6 D x:
6
6 _0
4 DE
xo I
M1 D
0
0
0
M1 K2
T1
d0 K3
T1
A K5 KA
32
3 2
3
0
Dd
0
7 6
7
6
0 7
76 Dx0 7 6 0 7Dupss
54 DEq 5 4 0 5
T1
d0
DE0fd
T1
T1
A
A KA
5:30
where
#
0
DE0q1
DEfd1
0
Dd
; Dx
;
;
;
DE
fd
0
DEq2
DE0fd2
Dd2
Dx2

k11 k11
k211 k212
k311 k312
K1
; K2
; K3
;
k12 k12
k221 k222
k321 k322

k41 k41
k51
k611 k612
k51
; K5
; K6
;
K4
k42 k42
k52 k522
k621 k622

0

M1 0
Tdo1
D1 0
0
M
;D
; Td0
;
0
T0do2
0 M2
0 D2

Dupss1
KA1
TA1
0
0
KA
; TA
; Dupss
Dupss2
0
KA2
0
TA2
Dd1
Dx1
"
DE0q
The model can be shown in Fig. 5.5 which is the matrix form of Fig. 5.4.
192
2
k11
E q 2 '
k 212
k11
1
D1 + sM1
k 211
0
s
1
E q 2 '
k 312
k 41
k 41
k 51
k 51
E q1 '
1
k 311 + sTd01 '
-+
+
E fd1 '
K A1
1 + sTA1
u pss1
k 612
K 611
E q 2 '
k1 2
E q1 '
k1 2
k 2 21
1
D 2 + sM 2
k 2 22
0
s
E q1 '
k 3 21
k 4 2
k 5 2
k 4 2
k 5 2
E q 2 '
1
k 3 22 + sTd02 '
-+
+
E fd 2 '
KA2
1 + sTA 2
k 6 22
u pss2
k 6 21
E q1 '
Fig. 5.4 HeffronPhillips model of two-machine power system installed with PSS
5.1.2
A Multi-machine Power System Installed with Power

System Stabilizers
5.1.2.1
HeffronPhillips Model of an N-Machine Power System

Installed with Power System Stabilizers
In an N-machine power system installed with PSSs, the 4th-order model of synchronous generator is given by Eqs. (5.24) and (5.25) for i 1; 2; . . .; N.
Equation (5.8) for i 1; 2; . . .; N can be written in the following matrix form
193
K1
0 I
s
(sM + D) 1
K4
K5
K2
(K 3 + sTd0 )
(I + sTA ) K A
u pss
K6
Fig. 5.5 HeffronPhillips model of two-machine power system in matrix form
g ejd E0 jx0 Ig x x0 ejd90 iq

V
q
d
d
q
5:31
where

g V
g2 . . . V
gN T ; Ig Ig1 Ig2 . . . IgN T ;
g1 V
V

T
T
E0q E0q1 E0q2 . . . E0q1 ; iq iq1 iq2 . . . iq1 ;
ejd diagejdi ; x0d diagx0di ; xq x0d ejd90 diagxqi x0di ejdi 90
5:32
In Eq. (5.32), diag(ci denotes an Nth-order diagonal matrix with ci being the ith
diagonal element.
Let the network equation be
Ig Y
g
NV
5:33
N is the network admittance matrix with only the nodes of generator

where Y
terminal left. From Eqs. (5.31) and (5.33), it can have
h
i
Ig Y
ejd E0 xq x0 ejd90 iq Y
E g
q
d
5:34
Y
1 jx0 1 . The above equation gives the generator current in xy
where Y
d
N
coordinate as (i 1; 2; . . .; N)
194
Igi
N
X
5:35
j1
Hence, Eq. (5.15) can be obtained

where
yij yij ejaij is the elements of matrix Y.
for i 1; 2; . . .; N. Their linearization can be expressed in the following matrix
form
DId Fdd Dd Gdd DE0q Hdd DIq
5:36
DIq Fqq Dd Gqq DE0q Hqq DIq

where
Dd Dd1
Dd2

. . . DdN T ; DE0q DE0q1
DId Did1
Did2
...
DidN T ; DIq Diq1
DE0q2
Diq2
DE0qN
...
...
DiqN T
f ddij yij E0qj0 cosdj0 di0 aij xqj x0dj iqj0 sindj0 di0 aij ; i 6 j,
f qqij yij E0qj0 sindj0 di0 aij xqj x0dj iqj0 cosdj0 di0 aij ; i 6 j,
f ddii
N
X
j1;j6i
f ddij ; f qqii
N
X
f qqij ;
j1;j6i
gddij yij sindj0 di0 aij ; gqqij yij cosdj0 di0 aij ;
hddij yij xqj x0dj cosdj0 di0 aij ; hqqij yij xqj x0dj sindj0 di0 aij
and f ddij ; f qqij ; gddij ; gqqij ; hddij and hqqij are the elements of matrix Fdd ; Fqq ; Gdd ;
Gqq ; Hdd and Hqq , respectively.
DId Fd Dd Gd DE0q
DIq Fq Dd Gq DE0q
where
Fd Fdd Hdd Fq ;
1
Fq I Hqq Fqq ;
Gd Gdd Hdd Gq ;
Gq I Hqq 1 Gqq
5:37
195
Linearization of Eqs. (5.24) and (5.25) is

Dd_ i xo Dxi
1
Dx_ i
DPi Di Dxi
Mi
1
0
DE_ qi 0 DEqi DE0fdi
Tdoi
1
KAi
0
DE_ fdi
DE0
DVgi Dupssi
TAi fdi TAi
5:38
Dvdi xqi Diqi

Dvqi DE0qi x0di Didi
DPi vdi0 Diid vqi0 Diqi idi0 Dvdi iqi0 Dvqi
DEqi DE0qi xdi x0di Didi
vgdi0
vqi0
DVgi
Dvgdi
Dvgqi
Vgi0
Vgi0
5:39
In matrix form, the above equations are

Dd_ xo IDx
Dx_ M1 DP DDx
5:40
0
DE_ q T01
d0 DEq DEfd
0
0
1
DE_ fd T1
A DEfd TA KA DVg Dupss
DP Iq0 DE0q Vd0 Iq0 X0d DId Vq0 Id0 Xq DIq

DEq DE0q Xd X0d DId
DVg
0
V01
g0 Vgq0 DEq
5:41
0
V01
g0 Vgq0 Xd DId
V01
g0 Vgd0 Xq DIq
where
Dx Dx1
Dx2
...

DxN T ; DE0fd DE0fd1
DP DP1
DP2
...
DPN T ; DEq DEq1
DE0fd2
DEq2
...
. . . DE0fdN
DEqN T ;
T
T
DVg DVg1 DVg2 . . . DVgN ; Dupss Dupss1 Dupss2 . . . DupssN ;
M diag(Mi ; D diag(Di ; T0d0 diag(T0d0i ; TA diag(TAi ; KA diag(KAi ;
Id0 diag(idi0 ; Iq0 diag(iqi0 ; Vd0 diag(vdi0 ; Vq0 diag(vqi0 ; Vg0 diag(Vgi0 ;
X0d diag(x0di ; Xq diag(xqi ; Xd diag(xdi
196

DPK1 Dd K2 DE0q
DEq K3 DE0q K4 Dd
5:42
DVg K5 Dd K6 DE0q
By substituting Eq. (5.42) into Eq. (5.40), HeffronPhillips model of the
N-machine power system with PSS installed can be obtained as
2
Dd_
6 Dx_
6
6 DE_ 0
4
q
0
DE_
fd
2
0
7 6
7 6 M1 K1
74
T1
5
d0 K4
T1
A K5 KA
xo I
M1 D
0
0
32
3 2
3
Dd
0
0
7 6
7
6
0 7
76 Dx0 7 6 0 7Dupss
54 DEq 5 4 0 5
T1
d0
DE0fd
T1
T1
A
A KA
0
M1 K2
T1
d0 K3
T1
A K5 KA
5:43
It is in the exactly same format to that of two-machine power system given by
Eq. (5.30). Hence, the model can also be shown in Fig. 5.5.
5.1.2.2
Linearized Model When Full Mathematical Model

of Synchronous Generators Is Used
Linearized full dynamic mathematical model of the ith synchronous generator in the
N-machine power system is (i 1; 2; . . .; N)
Dw_ di Dvdi rai Didi x0 Dwqi wqi0 Dxi
Dw_ qi Dvqi rai Diqi x0 Dwdi wdi0 Dxi
Dw_ fi Dvfi rfi Difi
5:44
Dw_ Di rDi DiDi

Dw_ Qi rQi DiQi
2
Didi
xdi
7 6
6
4 Difi 5 4 xadi
"
DiDi
Diqi
DiQi
"
xadi
xfi
xadi
31 2
Dwdi
7 6
7
xadi 5 4 Dwfi 5
xqi
xadi xDi
DwDi
#1 "
#
Dwqi
xaqi
xaqi
xQi
xadi
DwQi
5:45
197
Dd_ i x0 Dxi
1
Dx_ i DTei Di Dxi
Mi
DTei wdi0 Diqi wqi0 Didi iqi0 Dwdi idi0 Dwqi
5:46
1
KAi
Dv0
DVgi Dupssi
TAi fi TAi
vgdi0
vgqi0
Dvgdi
Dvgqi
DVgi
Vgi0
Vgi0
5:47
Dv_ 0fi
In matrix form, Eqs. (5.44) to (5.47) can be arranged as

X_ g Ag Xg Bg Dupss Bgv DVdq
5:48
DIdq Cg Xg
where

Dwd Dwd1
Dwf Dwf1

DwQ DwQ1
h
Xg DdT

T
T
. . . DwdN ; Dwq Dwq1 Dwq2 . . . DwqN ;
Dwf2 . . . DwfN T ; DwD DwD1 DwD2 . . . DwDN T ;
T
DwQ2 . . . DwQN ; Dv0f Dv0f1 Dv0f2 . . . Dv0fN T ;
iT
DxT DwTd DwTq DwTf DwTD DwTQ Dv0T
f
Dwd2
Dupss Dupss1
Dupss2
...
DupssN T
DVdq Dvgd1
Dvgq1
Dvgd2
Dvgq2
DIdq Did1
Diq1
Did2
Diq2
... ...
... ...
DvgdN
DidN
DvgqN T
DiqN T
Conversion equations between xy and di qi coordinate systems as shown in

Fig. 5.2 are (i 1; 2; . . .; N)

f di
sin di
f qi
cos di
cos di
sin di

f
f xi
sin di
; xi
f yi
f yi
cos di
cos di
sin di
f di
f qi

5:49
where f i is a phasor (i 1; 2; . . .; N). Linearization of above equations is

Didi
ixi0 cos di0 iyi0 sin di0
sin di0 cos di0 Dixi
Ddi
Diqi
Diyi
ixi0 sin di0 iyi0 cos di0
cos di0
sin di0

Dvgdi
vgxi0 cos di0 vgyi0 sin di0
sin di0 cos di0 Dvgxi
Ddi
Dvgqi
Dvgyi
vgxi0 sin di0 vgyi0 cos di0
cos di0
sin di0
5:50
198
In matrix form, above equation can be written as

DIdq Tg0 DIxy BgI Xg
DVdq Tg0 DVxy BgV Xg
5:51
where
DVxy Dvgx1
DIxy Dix1
Dvgy1
Diy1
Dvgx2
Dix2
Dvgy2
Diy2
...
...
DvgyN T
. . . DvgxN
DiyN T
. . . DixN
In the xy coordinate, network equation of Eq. (5.33) can be written as

3 2

ix1
gN11 bN11
6 iy1 7 6
7 6 bN11 gN11
6
7
6

6 ix2 7 6
gN21 bN21
7 6
6
6
6 iy2 7 6 bN21 g
N21
7
6
6 . 7 6
.
6 .. 7 6
..
7 6
6

7 6
6
4 ixN 5 4 gNN1 bNN1
bNN1 gNN1
iyN
2
gN12
b
N12
gN22
bN22
gNN2
bNN2

bN12
gN12

bN22
gN22
..
.

bNN2
gNN2

...
...
..
.
gN1N
b
N1N
gN2N
bN2N
...
gNNN
bNNN
32 v 3
gx1
bN1N
6 vgy1 7
7
6
gN1N 7
7
6
7
7
v
7
6
gx2
bN2N 76
76 v 7
gN2N 76 gy2 7
7
76
76 .. 7
..
7
. 7
7
.
76
6
bNNN 54 vgxN 7
5
gNNN
vgyN
5:52
N.
where
yNij gNij jbNij is the elements of network admittance matrix Y
DIxy YG DVxy
5:53
where
2
gN11
b
N11
gN21
bN21
6
6
6
6
6
YG 6
6
6
6
6
4 gNN1
bNN1

bN11
gN11

bN21
gN21
..
.

bNN1
gNN1
gN12
b
N12
gN22
bN22
gNN2
bNN2

bN12

gN12

bN22

gN22
..
..
.
.
bNN2

gNN2
gN1N
b
N1N
gN2N
bN2N
gNNN
bNNN
3
bN1N
gN1N 7
7
bN2N 7
7
7
gN2N 7
7
7
..
7
.
7
bNNN 5
gNNN

X_ g Ag Xg Bg Dupss Bgv Tg0 DVxy BgV Xg
Cg Xg Tg0 YG DVxy BgI Xg
5:54
199
Hence, state equation of the power system is

X_ g AXg Bg Dupss
5:55
where
A Ag Bgv Tg0 Tg0 YG 1 Cg BgI BgV Xg
5.2
Modal Analysis and Control of Power System

Oscillations in a Multi-machine Power System
Installed with Power System Stabilizers
5.2.1
Eigensolution for the Analysis of Power System

Oscillations
5.2.1.1
Participation Factor, Correlation Ratio of Electromechanical

Loop, Modal Shape, and Eigensolution
Modal analysis is a method to study power system small-signal angular stability by

calculating various eigenparameters of state matrix of state-space representation of
power system linearized model of Eqs. (5.43) and (5.55).
In Sect. 2.2.1.1, following denition on eigenvalues ki , corresponding left and
right eigenvectors, vi and wTi , of state matrix A is given
Avi ki vi ;
wTi A wTi ki ;
i 1; 2; . . .; M
5:56
Also following two matrices are constructed by eigenvectors as

2
V v1
v2
. . . vn ; V1
3
wT1
6 wT 7
6 27
T
7
6
6 .. 7 W
4 . 5
5:57
wTn
Equations (2.61) and (2.66) establish the following relationship between the state
variables xi t; i 1; 2; . . .; M and modes of the system zi t zi 0eki t ; i
1; 2; . . .; M to be
xk t) vk1 z1 0ek1 t vk2 z2 0ek2 t vkn zn 0ekM t
5:58
200
zi t) w1i x1 t) w2i x2 t) wMi xM t)
5:59
where vki is the kth-row ith-column element of matrix V and wki the ith-row
kth-column element of matrix W.
From Eq. (5.58), it can be seen that the magnitude of vki measures how much the
ith mode zi t) zi 0eki t ; i 1; 2; . . .; M, contributes to the kth state variable
xk t). Thus, jvki j is a kind of measurement of observability of the ith mode in the
kth state variable. Equation (5.59) indicates that the magnitude of jwki j measures the
influence of the kth state variable xk t) on the ith mode zi t) zi 0eki t ; i
1; 2; . . .; M of the system. It is a kind of measurement of controllability of the kth
state variable on the ith mode. Hence, jvki wki j measures how much the ith mode and
the kth state variable are connected. Its normalized value is dened to be the
participation factor
jvki wki j
pki PM
i1 jvki wki j
5:60
If it is assumed that xk 0 1; xj 0 0 (j 6 k, j 1; 2; . . .; n), from

Eq. (5.59), it can have
zi 0 wki ; i 1; 2; . . .; M
5:61
By substituting Eq. (5.61) into Eq. (5.58) it can have

xk t)
n
X
vki wki eki t
5:62
i1
Equation (5.62) indicates that the participation factor jvki wki j measures how
much the ith mode excited by the kth state variable participates the time response of
the kth state variable. That is why the index dened by Eq. (5.60) is named as the
participation factor.
As it is pointed out previously in Chap. 2, the electromechanical power oscillations are closely related to the rotor motion equation of generators where state
variables are the deviation of rotor position Ddi and speed Dxi ; i 1; 2; . . .; M in
i ,
the state Eqs. (5.43) or (5.55). Hence for a pair of conjugate complex eigenvalue, k
of state matrix, if the associated participation factors of state variables Ddi and
Dxi ; i 1; 2; . . .; M are much higher than those of other state variables, this means
i is the dominant oscillation mode responsible for the electromechanical oscillak
i is an electromechanical oscillation mode. That is, if
tion. Hence, k
5.2 Modal Analysis and Control of Power System
X
p xk Dd pki xk Dx
1;
ri Pki
pki 1
k pki xk 6 Dd and xk 6 Dx
k
201
!
5:63
i is an electromechanical oscillation mode. ri dened by Eq. (5.63) is called the

k
correlation ratio of electromechanical loop, which measure how much closely the
i is related to the rotor motion of generators (hence the electromeeigenvalue k
i is an
chanical oscillation). ri is often used to identify whether an eigenvalue k
oscillation mode or not.
i , let its corresponding (right) eigenvector
For the oscillation mode of interests, k
T
be
vi
v1i
v2i . . . vMi . Assume the rotor speed of generators, Dxj
(j 1; 2; . . .; N), to be the (N + 1)th to (2N)th state variables of the system of
Eq. (4.1), i.e.,
Dxj Dxj N ;
j 1; 2; . . .; N
5:64

3
3
2
x1 t)
v1i
6 .. 7
6 .. 7
6 . 7
6 . 7
7
7
6
6
6 Dx1 7
6 vN 1i 7
7
7
6
6
6 Dx2 7
6 vN 2i 7

7
7
6
6
k1 t
6 .. 7 v1 z1 0e 6 .. 7zi 0eki t
6 . 7
6 . 7
7
7
6
6
6 DxN 7
6 v2Ni 7
7
7
6
6
6 .. 7
6 .. 7
4 . 5
4 . 5
vMi
xM t)
2
5:65
The (N + 1)th to (2N)th elements of vi , vki vki \uki ; k = N + 1, N +

i , because vki vki \u
2; . . .; 2N describe the mode shape of oscillation mode k
ki
weighs the contribution of the oscillation mode to the rotor motion of involved
generators. By sketching the phasor diagram of vki vki \uki , relative positions of
rotor motion of involved generators can be found to determine how generators
participate the power oscillation. For example, let
i n jxi ; vN 1i jvN 1i j\0 ; vN 2i jvN 2i j\180
k
i
5:66
Equations (5.65) and (5.66) indicate that the rotor motion of the 1st and 2nd
generator, Dx1 and Dx2 , must have component jvN 1i jzi 0eni cos xi t and
jvN 2i jzi 0eni cosxi t 180 , respectively. Hence, as far as oscillation mode
i n jxi is concerned, the 1st and 2nd generators oscillate against each
k
i
other. Thus, by looking at vki vki \uki ; k N 1; N 2;. . .; 2N, how groups
202
of generators take part in the oscillation associated with the mode can be
determined.
Procedure of modal analysis from the eigensolution of power system oscillations
normally is
1. Calculation of eigenvalues of state matrix;
2. Identication of electromechanical oscillation modes by checking the imaginary
part of eigenvalues within the range 2pf o f o 0:1 2 Hz and computing the
correlation ratio of electromechanical loop;
3. Examining how much each generator is involved in a particular oscillation mode
by calculating the modal participation index;
4. Determining how groups of generators oscillate against each other by computing
the modal shape.
5.2.1.2
Selective Reduced-Order Calculation of Electromechanical

Oscillation Modes
Commonly used numerical method of eigensolution is the QR decomposition. It is a

method to compute all eigenvalues of system state matrix. From above introduction
on the eigensolution for the analysis of power system oscillations, it can be seen
that in fact the majority of computational results of the QR method is redundant,
because only eigenvalues and associated eigenvectors related to the electromechanical oscillation modes are needed. In addition, for a large-scale power system
with a high-dimensional state matrix, numerical difculty of QR method may occur.
Hence, a lot of effort has been spent by many researchers to propose and develop
numerical methods for the efcient and effective eigensolution of large-scale power
system. The reduced-order modal analysis is one of the methods to carry out the
eigensolution from the reduced-order matrices rather than directly from the system
high-dimensional state matrix so as to avoid the numerical difculty. The selective
modal analysis is another computational strategy to compute only the eigenvalues
of interests from system state matrix to save computational time and resources. In
this section, a selective reduced-order method of modal computation is introduced
by using HeffronPhillips model.
HeffronPhillips model of N-machine power system of Eq. (5.43) can be written
as (without PSS being considered)
sDd xo Dx
sDx M1 K1 Dd K2 DE0q DDx
0
0
sDE0q T01
d0 K3 DEq K4 Dd DEfd
0
1
0
sDE0fd T1
A DEfd TA KA K5 Dd K6 DEq
5:67
203

fs2 M sD xo K1 K2 I sTA K3 sT0d0 KA K6 1
I sTA K4 KA K5 gDd FsDd 0
5:68
is an electromechanical oscillation mode, it must have

Obviously, if k
i
i 0
DETFk
5:69
where DETM 0 denotes the determinant of matrix M being equal to zero.

Usually, the oscillation modes of interests are poorly damped, close to the
imaginary axis of the complex plan. The angular oscillation frequency of electromechanical oscillation modes is close to the natural oscillation frequency, which
can be determined from the computation of eigenvalues of matrix xo M1 K1 . In
fact, in the N-machine power system, linearized 2N rotor motion equations of
generators are (with D being ignored)

0
Dd_
K1
Dx_
xo I
0
Dd
Dx
The characteristic equation of the system is

kI 0 xo I jkI Ar j 0

K1
0
5:70
5:71
The equivalent form of Eq. (5.71) is

s2 M xo K1 Dd s2 I xo M1 K1 Dd 0
5:72
Hence, 2N eigenvalues of system state matrix Ar are also solutions of the

following equation
2

s I xo M1 K1 0
5:73
Therefore, if N eigenvalues of matrix xo M1 K1 are gj ; j 1; 2;. . .; N, 2N

p
eigenvalues of matrix Ar are gj ; j 1; 2; . . .; N. It can be proved that all
eigenvalues of xo M1 K1 are real negative numbers except one zero eigenvalue.
Hence, 2N 2 eigenvalues of Ar are N 1 pairs of the conjugate imaginary roots
p
j gj ; j 1; 2; . . .; N. They are in fact the N 1 electromechanical oscillation
modes of the power system when only the rotor motion of generators is considered.
p
gj 6 0; j 1; 2; . . .; N 1; are the N 1 angular natural oscillation frequency
of electromechanical oscillation modes.
i n jxi , because it is
For a poorly damped oscillation mode of interests, k
i
close to the imaginary axis of the complex plan and the angular oscillation
204
Fig. 5.6 Optimal direct
one-dimensional searching
along the direction of
imaginary and real axis on
complex plane
Contour of f ( )
0 + j i
p
frequency xi is close to the angular natural oscillation frequency, gi , k
i 0
p
i n jxi .
0 j gi should be a good initial guess of oscillation mode, k
i
Therefore, following objective function can be established
fk jDETFkj
5:74
A direct searching method can be used to nd the minimal optimum of above

p
objective function on the complex plan, starting from 0 j gi . That will be the
i n jxi , which satises Eq. (5.74). The direct
oscillation mode of interests, k
i
searching can be implemented in the direction of imaginary and real axis iteratively

i n jxi is very close to 0 jp
as shown in Fig. 5.6. If k
gi , k
i ni jxi will
i
be found by applying one-dimensional optimal searching just a couple of times in
each of two directions.
5.2.2
Design of Power System Stabilizers

in Multi-machine Power System
5.2.2.1
Coordinated Design of Multiple Power System Stabilizers
In an N-machine power system, it is assumed that there are L oscillation modes of

interests which need to be damped by installing PSSs in the power system. The rst
task of designing PSSs is to determine the effective installing locations of PSSs.
Though nowadays it may be the case that every generator is equipped with a PSS, it
is still necessary to determine how the PSS is designed to effectively damp a
205
particular oscillation mode. This is an equivalent task to selecting the installing

locations of the PSS.
By computing the participation factors and/or the correlation ratio of electromechanical loop, it can be decided which generators are most strongly related to
a particular oscillation mode of interests. This is one of the most commonly used
methods to select the installing locations of PSSs to damp the particular oscillation
mode of interests. Without loss of generality, it can be assumed that L generators in
the N-machine power system are selected as the most effective installing locations
of PSSs to damp L oscillation modes of interests. The ith oscillation mode of
i n jxi is selected to be damped by the PSS installed on the jth
interests k
i
generator in the power system, denoted as the jth PSS. In total, L PSSs are to be
designed to damp L oscillation modes. (It may be possible that more than L PSSs
are needed to damp L oscillation modes in some cases.)
A simple way to design L PSSs is to set parameters of L PSSs one by one in a
sequence. For example, method of pole assignment introduced in Sect. 2.2.2.3 can
i n jxi to a target position k

be used to assign the ith oscillation mode k
i
i
n i jx i by setting the parameters of the jth PSS. Afterwards, the (i + 1)th oscillation mode is assigned by designing the (j + 1)th PSS, etc., until L oscillation
modes are assigned to the target positions by setting parameters of L PSSs.
However, this method of sequential design of L PSSs has a problem of eigenvalue
drifting, because when the (j + 1)th PSS is designed to assign the (i + 1)th
oscillation mode, the ith oscillation mode which has been assigned may drift away
from the target position. Hence at the end, L PSSs may not assign L oscillation
modes to their target positions at all.
To solve the problem of eigenvalue drift associated with the method of
sequential design of multiple PSSs, parameters of L PSSs can be set to move L
oscillation modes to the target positions simultaneously. This strategy of simultaneous tuning of parameters of all PSSs is also referred to as the coordinated design
of multiple PSSs. So far, there have been many methods of simultaneous tuning or
coordinated design of PSSs being proposed and developed. The majority of those
proposed methods adopt a certain objective function,
fp; p 2 P
5:75
where p is the vector of parameters of L PSSs and P is the parameter space of L

PSSs. A method of optimization can be used to nd the optimum of objective
function, which is the solution of coordinated design of L PSSs.
So far, methods of optimization for the coordinated design of multiple PSSs can
be classied into two categories: (1) mathematical methods and (2) heuristic
methods. Mathematical methods include linear programming, nonlinear programming and other mathematical algorithms of optimization. Examples of heuristic
methods are genetic algorithm, simulated annealing, and particle swarm algorithm.
Those methods have all been tried for optimally tuning parameters of multiple PSSs
to effectively damp power system oscillations. Application of mathematical methods
206
must have a computable objective function and sometimes is constrained by the

problem of local optimum, while heuristic methods can handle non-computable
objective function and are more effective to nd the global optima.
In the following section, a mathematical method is introduced to tune parameters
of multiple PSSs installed in a multi-machine power system on the basis of
HeffronPhillips model. In Sect. 7.4.1, a heuristic method to tune parameters of
multiple stabilizers added on a UPFC will be introduced.
5.2.2.2
Parameter Tuning Algorithm for the Design of Multiple Power

System Stabilizers
Let the objective function to simultaneously tune parameters of L PSSs to assign L

i n jxi , to target positions k
n jx , i 1; 2; . . .; L be
oscillation modes, k
i
i
i
i
fp
L
X
i pk
p 2
k
i
5:76
i1
The minimal optimum p 2 P is when the objective function is equal to zero and
L oscillation modes are moved to the target positions. If the deepest decent method
in nonlinear programming is used, the optimal searching algorithm is
pk 1 pk st rfpk
5:77
where st is the optimal one-dimensional searching length and rfpk is the

searching decent direction of objective function at the (k + 1)th step. A selective
modal algorithm to implement the deepest decent searching can be developed as
follows.
HeffronPhillips model of N-machine power system of Eq. (5.43) can be written
as (with PSSs being considered)
sDd xo IDx
0
0
sDE0q T01
5:78
0
1
0
sDE0fd T1
A DEfd TA KA K5 Dd K6 DEq Dupss

I sTA K4 KA K5 gDd K2 I sTA K3 sT0d0

1
KA K6 KA Dupss 0
5:79
207
Let the transfer function of the jth PSS be Gj s and the feedback signal Dxj .
Dupss can be denoted as
Dupss
s
Gpss p; sDd
x0
5:80
where Gpss p; s diagGpssj , Gpssj Gj s) when the jth generator is installed

with a PSS; otherwise, Gpssj 0. From Eqs. (5.79) and (5.80), it can have
s
I sTA K4 KA K5 KA Gpss p; sDdgDd Fp; sDd 0
x0
5:81
Hence, it must have

DETFp; ki 0
5:82
i
Equation (5.82) indicates that there should exist a nonzero vector vi and w
satisfying
Fp; ki vi 0
5:83
wTi Fp; ki 0
Let p p1 p2 . . . ps T . A small increment of pk , Dpk , will cause that of ki

and
vi , Dki and Dvi . It can have
Fppk Dpk ; ki Dki vi Dvi 0
5:84
Because
@Fp; ki
@Fp; ki
Dki
Dpk
@ki
@pk
Fppk Dpk ; ki Dki Fp; ki
5:85

i
Ti @Fp;k

w
@ki
@ pk vi

@Fp;k
i
@pk
T
vi
w
5:86
@ki

rfpk 2ki k i
@ki
@pk
5:87
208
where
@ki
@pk
@ki
@ p1
@ki
@ p2

@ki
@ ps
iT
i
and @@k
pk (k 1; 2; . . .; s) is calculated by
using Eq. (5.86).

When the deepest decent algorithm of Eqs. (5.77) and (5.87) is implemented, it
needs to compute the oscillation mode of interests, ki , at each step of optimal
searching. Because it is an iterative algorithm, at the (k + 1) step, ki k 1 is very
close to ki k and it should have DETFpk 1; ki k 1 0. Hence, set up the
following objective function,
f k 1 ki jDETFp; ki j
5:88
The direct searching method introduced based on Eq. (5.74) can be used to nd
ki k 1 on the complex plane, starting from ki k.
In the multi-machine power system, objective of installation of multiple PSSs is
to ensure sufcient damping to be provided to the oscillation modes of interests.
Hence, pole assignment to the exact target positions, such as that to be achieved by
the objective function of Eq. (5.76), is usually not necessary. A more relaxed
objective function can be used for the coordinated design of multiple PSSs. For
example, following objective function
fp minfi
0:05
5:89
where
ki ni jxi 2 fall oscillation modes to be damped by the PSS designg;
fi
fi q
2
ni x2i
is for the coordinated design of multiple PSSs to ensure the damping of all oscillation modes to be greater than .05. For the objective function of Eq. (5.89), some
heuristic methods of optimization, such as the genetic algorithm, can be used.
Strategies of sequential setting and simultaneous tuning for the design of multiple PSSs can also be jointly used to achieve more efcient and practical design.
For example, it can be assumed that in the N-machine power system, the jth PSS
provides a pure damping torque, Dpssj Dxj (j 1; 2; . . .; L), to the electromechanical
oscillation loop of the jth generator. With this assumption, Eq. (5.81) becomes
fs2 M sD Dpss xo K1 K2 I sTA K3 sT0d0 KA K6 1
I sTA K4 KA K5 gDd Fp; sDd 0
where
5:90
209
Dpss diag(dpssj ; dpssj Dpssj

when the jth generator is installed with a PSS; otherwise, dpssj 0 j 1; 2; . . .; L),

T
p Dpssj Dpssj 1 . . . Dpssj L .
The method of coordinated design of PSSs introduced from Eqs. (5.76) to (5.88)
above can be used to set the damping torque provision from each PSS for the
multiple PSSs to damp all the oscillation modes of interests. Afterwards, each PSS
can be designed individually by using the phase compensation method to ensure
that the right amount of damping torque, Dpssj Dxj (j 1; 2; . . .; L), is supplied to
each generator.
5.2.3
Fixed Modes Associated with PSS Control
A PSS is a local controller, and hence, an N-machine power system with L PSSs
installed forms a decentralized control system. Fixed modes are an important
concept about the controllability and observability of a decentralized control system, a natural generalization of the well-known concept of uncontrollable and
unobservable modes in centralized control systems. In this section, a mathematical
proof is presented to show that in the decentralized control of N-machine power
system by PSSs, any electromechanical oscillation mode is not a xed mode. This
means that in the N-machine power system, any oscillation mode can be damped by
the appropriate design of multiple PSSs. In other words, in the parameter space of
coordinated design of multiple PSSs, the optimum always exists if the objective
function is appropriately set. Theoretical proof given below is based on the
HeffronPhillips model of N-machine power system installed with multiple PSSs.
Let the realization of the jth PSS, Dupssj Gj sDxj , be
Dupssj Hj zj t kj Dxj
z_ j t Fj zj t Sj Dxj
5:91
Dupss Hzt Ky
z_ t Fzt Sy
5:92
Hence, it can have
where
210

T
T
zt z1 tT z2 tT zL tT ; y Dxj Dxj 1 Dxj L ;
H blockdiagHj ; F blockdiagFj ; S blockdiagSj ; K = diag(kj
and blockdiagMj denotes a block diagonal matrix with the diagonal matrix to be Mj .
From Eq. (5.92), state equation of HeffronPhillips model of N-machine power
system with L PSSs installed of Eq. (5.43) can be written as
_ Axt BDupss
xt
5:93
y Cxt
where
2
6
6 M1 K1
A6
6 T1 K
4
d0 4
xo I
M1 D
M1 K2
T1
d0 K3
T1
d0
T1
A K5 KA
T1
A
T1
0
A K5 KA
2
3
0
6
7
6 0 7
6
7; C 0 I
B6
7
4 0 5
3
7
7
7;
7
5
0
T1
A KA
The set of xed modes of the decentralized control system of Eqs. (5.92) and
(5.93), denoted as KA; B; C; K, is dened as
KA; B; C; K \ 0 kA BKC
K2K
5:94
where kA BKC denotes the set of all eigenvalues of matrix A BKC, and K0 is
the following set
K0 fKjK diagkj ; K 2 RNN g
5:95
Denition of the xed modes given by Eqs. (5.94) and (5.95) in fact means that
the xed modes are not affected by variations of any kj . This is the key in the
following proof.
0
6 M1 K1
A BKC 6
4 T1 K4
d0
T1
A K5 KA
0
xo I
M1 D
M1 K2
0
T1
d0 K3
1
1
TA KA K TA K5 KA
211
3
0
0 7
7
5
T1
d0
1
TA
5:96
If ko is a xed mode, it must be the eigenvalue of matrix A BKC. Hence,

32 3
2
2 3
0
0
0
xo I
v1
v1
1
1
1
7
7
7
6 M K1
6
6
M
D
M
K
0
v
2
76 2 7 ko 6 v2 7 5:97
6
1
1 54
5
4 T1 K4
4
0
Td0 K3
Td0
v3
v3 5
d0
1
1
1
1
TA K5 KA TA KA K TA K5 KA TA
v4
v4
That is,
xo v2 ko v1
K1 v1 Dv2 K2 v3 ko Mv2
K4 v1 K3 v3 v4 ko Td0 v3
K5 KA v1 KA Kv2 K6 KA v3 v4 ko TA v4
5:98

A4 k4o
A3 k3o
A2 k2o

KA K
A1
k o A 0 v1 0
xo
5:99
where
A0 K3 KA K6 K1
2 K1 K4 KA K5
1
A1 TA K3 Td0 K1
K KA K6 K1
2 K1
2 D TA K4
x0 3
1
1
1
A2
TA K3 Td0 K1
K KA K6 K1
2 D TA Td0 K2 K1
2 M
x0
x0 3
1
1
A3
TA Td0 K1
T K3 Td0 K1
2 D
2 M
x0
x0 A
1
A4
TA Td0 K1
2 M
x0
v 6 0; hence, v1 6 0. It should have

KA K
DET A4 k4o A3 k3o A2 k2o A1
ko A0 0
xo
Denote
5:100
212
Fko ; k1 ; k2; . . .kN
A4 k4o
A3 k3o
f 11 ko ; k1
6 f 21 ko
6
6
..
6
4
.
f N1 ko

KA K
A1
ko A0
xo
3
f 12 ko
...
f 1N ko
f 22 ko ; k2 . . .
f 2N ko 7
7
7
..
..
..
7
5
.
.
.
A2 k2o
f N2 ko
...
5:101
f NN ko ; kN
Also denote
2
f 11 ko ; k1
6 f 21 ko
6
Fk ko ; k1 ; k2 ; . . .kN 6
..
4
.
f 12 ko
...
f 22 ko ; k2 . . .
..
..
.
.
f N2 ko
...
f N1 ko
f 1N ko
f 2N ko
..
.
3
7
7
7 5:102
5
f NN ko ; kN
Obviously, FN ko ; k1 ; k2 ; . . .kN Fko ; k1 ; k2 ; . . .kN . Let the companion

matrix of Fk ko ; k1 ; k2; . . .kk be F k ko ; k1 ; k2; . . .kk and denote
2
f k;11
6 f k;21
6

Fk ko ; k1 ; k2; . . .kk 6 .
4 ..
f k;k1
f k;12
f k;22
..
.
f k;k2
...
...
..
.
...
3
f k;1k

f k;2k 7
7
.. 7
. 5
f k;kk
5:103

f k;kk DETFk1 ko ; k1 ; k2; . . .kk1
5:104
gk ko ; k1 ; k2; kk DETFk ko ; k1 ; k2; kk
5:105
Denote
When k N, from Eqs. (5.100) and (5.105) it can have

gN ko ; k1 ; k2 ; . . .kN DETFN ko ; k1 ; k2 ; . . .kN
DETFko ; k1 ; k2 ; . . .kN 0
5:106
Because ko is a xed mode,

@ko
0j 1; 2; . . .; N
@kj
5:107
213
@gN @ko
@g
N0
@ko @kN @kN
5:108

@gN
0
@kN
5:109
dDETMx
dMxT
TraceM x

dx
dx
5:110
For any matrix Mx,
where M x is the companion matrix of Mx; thus,

@gN
@FN ko ; k1 ; k2 ; . . .kN T
TraceF N ko ; k1 ; k2 ; . . .kN
0
@kN
@kN
5:111

h
@FN ko ; k1 ; k2 ; kN T
diag 0
@kN
h
diag 0
0
@ f NN ko ;kN
@ kN
0
ko
x0
i
5:112
KNA
Hence from Eqs. (5.103), (5.111), and (5.112), it can have

@gN ko
KNA f N;NN 0
@kN x0
5:113
KNA 6 0 and if ko is an electromechanical oscillation mode, ko 6 0. Hence from

Eqs. (5.113), (5.104), and (5.105), it can have
f N;NN DETFN1 ko ; k1 ; k2 ; . . .kN1 gN1 ko ; k1 ; k2 ; . . .kN1 0 5:114
From Eq. (5.114), it should have
@gN1 @ko
@g
N1 0
@ko @kN1 @kN1
From Eqs. (5.101), (5.107), and (5.110), it can be obtained
5:115
214
"
#
@gN1
@FN ko ; k1 ; k2 ; . . .kN1 T

Trace FN1 ko ; k1 ; k2 ; . . .kN1
@kN1
@kN1
n
h
io
Trace F N1 ko ; k1 ; k2 ; . . .kN1 diag 0 0 . . . kko0 KN1A
@o
@o
KN1A f N1;N1N1 KN1A gN2 @o ; k1 ; k2 ; . . .kN2 0
@0
@0
5:116
Again, KN1A 6 0 and ko 6 0, it can have

gN2 ko ; k1 ; k2 ; . . .kN2 DETFN2 ko ; k1 ; k2 ; . . .kN2 0
5:117
Above procedure from Eqs. (5.106) to (5.117) can be carried on iteratively until
it is obtained that
@g1 ko
K1A 0
@k1 x0
5:118
It contradicts to that K1A 6 0 and ko 6 0. Hence, ko cannot be a xed mode and

an electromechanical oscillation mode.
5.3
An Example Three-Machine Power System
5.3.1
Example Power System and Its Linearized

5.3.1.1
System Parameters and Operating Conditions
Conguration of a three-machine six-node example power system is shown in

Fig. 5.7 [1]. Parameters of generators and the AVRs are given in Table 5.1.
Parameters of transformers and lines are given in Table 5.2[1]. Initial conditions of
example power system for the load flow computation are
VG1 1:04\0; VG2 1:025\0; VG3 1:025\0;
V1;2;3 1:00\0; VA;B;C 1:00\0;
SG1 0:716 j0:27; SG2 1:63 j0:067; SG3 0:85 j0:109
LoadA 1:25 j0:5; LoadB 0:9 j0:3; LoadC 1 j0:35:
Tables 5.3 and 5.4 are the results of the load flow computation.
5.3 An Example Three-Machine Power System

0.716+j0.27
BusC
Bus1
215
1.25 + j0.5
Bus3
1.63 + j0.067
G1
1.040
G3
1.0250
BusB
BusC
0.9 + j0.3
1 + j0.35
Bus2
0.85 j0.109
G2
1.0250
Fig. 5.7 Conguration of a three-machine example power system [1]
Table 5.1 Parameters of

generators and the AVRs [1,
2]
xd
x0d
T0d0
xq
x0q
M
D
KA
TA
5.3.1.2
G1
G2
G3
0.146
0.0608
8.96
0.8958
0.1189
6
1.313
0.1813
5.89
0.0969
0.0969
0.8645
0.1969
1.258
0.25
47.2
0
200
0.02
12.8
0
200
0.02
6.02
0
200
0.02
Calculation of Initial Values of State Variables
From active and reactive power output of each generator, Pi and Qi ; i = 1,2,3, and
gi ; i = 1,2,3, output current of each generator can be calcuits terminal voltage, V
lated as
216
Table 5.2 Parameters of transformers and lines [1]

Node
Node
Resistance
Reactance
Susceptance
1
1
A
B
2
C
G1
G2
G3
A
B
2
3
C
3
1
2
3
0.01
0.017
0.032
0.039
0.085
0.0119
0
0
0
0.085
0.092
0.161
0.17
0.072
0.1008
0.0576
0.0625
0.0586
0.088
0.079
0.153
0.170
0.0745
0.1045
0
0
0
Table 5.3 Results of load flow computationnode voltage

Node
Magnitude of voltage
Phase of voltage (deg.)
G1
G2
G3
1
A
B
C
2
3
1.04
1.025
1.025
1.018
0.977
1.007
1.012
1.021
1.030
0.000
0.163
0.081
0.039
0.070
0.065
0.012
0.065
0.034
Table 5.4 Results of load flow computationactive and reactive power

Starting
node (SN)
Ending
node (EN)
Active
power at
SN
Reactive
power at SN
Active
power at
EN
Reactive
power at EN
4
4
5
6
8
7
G1
G2
G3
5
6
8
9
7
9
4
8
9
0.412
0.306
0.841
0.596
0.765
0.240
0.718
1.630
0.850
0.411
0.036
0.117
0.061
0.018
0.204
0.396
0.142
0.072
0.409
0.304
0.865
0.609
0.760
0.241
0.718
1.630
0.850
0.468
0.108
0.084
0.065
0.054
0.103
0.360
0.017
0.112
217
Igi Pi jQi ;

V
i 1; 2; 3
gi
Qi ; i 1; 2; 3, can be computed as
The imaginary voltage of each generator, E
Qi Vgi jxqiIgi ;
E
i 1; 2; 3
Qi ; i 1; 2; 3 is the angular position of each generator, di , i 1; 2; 3.

Phase of E
From di , i 1; 2; 3, computational results of output current and terminal voltage of
each generator can be transformed from the common xy coordinate to the dq
coordinate as (Eq. (5.49))
idi
iqi
sindi
cosdi
cosdi
sindi
vgxi
vgyi
vdi
vqi
sindi
cosdi
cosdi
sindi
vgxi
vgyi
According to Eq. (5.25), it can have

E0qi vqi x0di idi
Eqi E0qi (xdi x0di idi
Since E0fdi0 0, from the third equation in Eq. (5.24) it can have
Efd0i = Eqi0
Computational results of all the above are presented in Table 5.5.
5.3.1.3
Linearized HeffronPhillips Model
From Table 5.2 and Fig. 5.7, following network admittance matrix can be
established.

L Y11
Y
21
Y
Table 5.5 Initial value of
state variables
12
Y
22
Y
G1
G2
G3
di0 (rad.)
xi0
E0qi0
0.0621
1
1.0637
1.0388
1
0.8119
0.9199
1
0.7882
E0fqi0
1.0998
1.8302
1.4334
218
where
2
11
Y
j17:3611
0
4
0
j16
0
0
2
12
Y
j17:3611
4
0
0
2
21
Y
j17:3611
6
0
6
6
0
6
6
6
0
6
4
0
0
3:3074 j39:3089
6 1:3652 j11:6041
6
6 1:9422 j10:5107
6
6
6
0
6
4
0
0
221
Y
222
Y
0 0
0 0
0 0
0
6
0
6
6
0
6
6
6 3:7482 j23:6449
6
4 1:6171 j13:6980
1:1551 j9:7843
0
0
0
0
0
0
0
j16
0
3
0
5;
0
j17:0648
0
j16
0
3
0
5;
0
j17:0648
3
0
7
0
7
7
0
7
7;
7
0
7
5
0
j17:0648
3
1:9422 j10:5107
7
0
7
4:1115 j16:1367 7
7
7;
7
0
7
5
0
1:2820 j5:5882
1:3652 j11:6041
3:8620 j18:0714
0
0
1:1876 j 5:9751
0
0
1:1876 j5:9751
0
1:6171 j13:6980
2:8047 j35:4456
0
3
0
7
0
7
1:2820 j5:5882 7
7
7;
1:1551 j9:7843 7
7
5
0
2:4371 j32:1539
22 Y
222
221 Y
Y
By keeping three nodes of generator terminals via deleting all the other nodes in
the above network admittance matrix, it can have
N Y
1 Y

11 Y
12 Y
Y
22 21
2
1:0994 j4:7578
6
4 0:0954 j2:2197
0:7374 j5:1374
7
0:1249 j2:8146 5
0:0057 j2:2568
0:1249 j2:8146
0:7236 j5:0289
0:0954 j2:2197
0:0057 j2:2568
219
From Eq. (5.34), it can be obtained

Y
1 jx0
Y
N
d
2
0:8312 j3:0501
6
4 0:2806 j1:4829
0:4192 j2:7499
7
0:2119 j1:0795 5
0:2054 j1:2048
0:2119 j1:0795
0:2759 j2:3757
3:1614\1:3047
1:5092\1:3838
1:2221\1:4019
6
4 1:5092\1:3838
1:2221\1:4019
0:2806 j1:4829
2:7816\1:4195
1:1001\1:3770
0:2054 j1:2048
7
1:1001\1:3770 5
2:3917\1:4552
Thus according to Eq. (5.37), it can have

2
0:1450
6
Fdd 4 1:4886
2:4386
0:0034
0:8120
1:9395
3:0501
6
4 0:5975
1:0626
0:9434
0:6326
0:0300
6
Hdd 4 0:0500
0:0377
2:7499
1:0970
0:7991
2:8414
6
7
0:9500 5; Fqq 4 0:6023
1:1275
Gdd
0:1484
7
1:0467 5; Gqq
2:3757
3
1:2064
1:2507
7
0:6042 5;
0:6478 0:8347 1:4825

2
3
0:8312 1:0718 0:7770
6
7
4 1:3859 0:4192
0:3384 5;
2
0:8366
1:5907
1:0457
0:0823
0:2759
0:1101
0:7923
1:0157
0:3126
6
7
0:3643 5; Hqq 4 0:0216
0:0614
0:2971
0:0228
2:0503
7
1:1270 5:
0:8179
2:5579
1:3566
0:8023
From Eq. (5.38), following matrices are calculated

2
0:3252
6
Fd 4 1:5477
0:0095
2:4103
1:1306
0:8935
2:2175
6
Gd 4 0:1557
1:2518
0:3995
1:1026
2:7758
0:3157
2:1589
7
6
0:8626 5; Fq 4 0:2670
0:3235
7
0:0565 5;
0:2296 0:1689 0:3985

3
2
1:1666
1:6110 0:8094 0:4956
7
7
6
0:9860 5; Gq 4 0:6318 0:1512
0:1474 5:
2:0241
2:3982
0:4495
0:0527
0:1083
220
Thus according to Eqs. (5.41) and (5.42), it can have

0
K1 Vd Iq Xd ) Fd Vq Id Xq Fq ;
0
K2 Iq Vd Iq Xd ) Gd Vq Id Xq Gq ;
0
K3 I (Xd Xd ) Gd ;
0
K4 Xd Xd ) Fd ;
0
1
K5 V1
gen Vq Xd Fd Vgen Vd Xq Fq ;
0
1
1
K6 V1
gen Vq Vgen Vd Xd Gd Vgen Vd Xq Gq
Coefcient matrices of HeffronPhillips model of example power system as

given by Eq. (5.43) are obtained to be
2
2:3373
6
K1 4 1:5290
2
1:4640
0:8201
1:8800
1:1889
0:1067
0:0994
1:2478
0:0068
6
K5 4 0:0596
0:0076
0:0780
0:0499
47:2
6
M4 0
2
Td0
200
8:96
6
4 0
0
0:1278
0:5628
0:6458
0:0277
7
0:4060 5;
3:0671
0:3694
0:4691
2:3234
0:0008
1:8725
200
0
0:0269
6
7
0 5; D 4 0
7
0 5;
0 0 0
6:02
3
2
0
0:02
0
7
6
0 5; TA 4 0
0:02
6:00
0
7
0 5:
5:89
7
0 5;
0:02
7
0:6701 5;
0:4790
200
0
0:1829
12:8
6
KA 4 0
0:9034
2:4574
3:7141
1:2795 1:0112 2:2907
2
3
3
0:0143
0:8732 0:0711 0:0678
6
7
7
0:0281 5; K6 4 0:4315 0:5292 0:1730 5:
0:0315
7
6
0:7660 5; K4 4 1:2024
3:1565
0:4521
7
6
0:6870 5; K2 4 1:0247
2:2160
1:0599
6
K3 4 0:1209
2
0:8733
5.3.2
Modal Analysis of Example Power System
5.3.2.1
Electromechanical Oscillation Modes of Example Power

System
221
Eigenvalues of state matrix of HeffronPhillips model obtained in the above section

are calculated as
1;2 25:0316 j28:2289
k
3 34:2648
k
4;5 25:0343 j9:5156
k
6 13:5354
k
7;8 1:5051 j9:1860
k
9;10 0:1736 j7:0590
k
11;12 0:0000 j0:0000
k
7;8 , and k
9;10 is 9:5156 1:5145 Hz, 9:1860 1:4620 Hz, and
4;5 , k
Frequency of k
2p
2p
1:1235 Hz; respectively, falling within the range of power system
low-frequency oscillation 0.12 Hz. They can be identied to be the electromechanical oscillation modes of the example power system. To conrm the identication, participation factor and the correlation ratio of electromechanical loop can
4 25:0343 j9:5156, k
7 1:5051 j9:1860, and k
9
be calculated for k
0:1736 j7:0590 as follows.
4 25:0343 j9:5156 are
Right and left eigenvectors corresponding to k
calculated as
7:0590
2p
3
2
3
0:6852 j17:3385
j0:0001
6 0:6595 j5:9033 7
6 0:0002 j0:0003 7
7
6
6
7
6 1:3447 j11:4352 7
6 0:0001 j0:0001 7
7
6
6
7
7
6 64:7503 j192:9710 7
6
0
7
6
6
7
7
6 31:8356 j61:9802 7
6
0
7
6
6
7
7
6 32:9147 130:9908 7
6
0
7
6
6
7
7 6
7 6
v
7
7; w
1:9362

396:4038
0:0020
j0:0008
7
6
6
7
6 0:0044 j0:0017 7
6 2:5524 j88:6491 7
7
6
7
6
6 0:0024 j0:0010 7
6 2:5028 j50:7603 7
7
6
7
6
6 0:5202 j0:0006 7
6 0:5973 j1:5444 7
7
6
7
6
4 0:7505 j0:0000 5
4 0:18208 j0:5224 5
0:12978 j0:2957
0:4076 j0:0030
2
By use of Eq. (5.60), participation factors corresponding to each of state variables are calculated as
222
p41 0:0017; p42 0:0022; p43 0:0018:

Thus, the correlation ratio of electromechanical loop can be calculated by using
4 25:0343
Eq. (5.63) as r4 0:0040, which is smaller than 1. Hence, k
j9:5156 is not an electromechanical oscillation mode.
7 1:5051 j9:1860 are calRight and left eigenvectors corresponding to k
culated as
3
2
3
1:0451 j2:2184
0:0007 j0:0011
6
6 0:0073 j0:0198 7
7
2:2567 j5:5000
7
6
6
7
6 3:3018 j7:7184 7
6 0:0252 j0:0502 7
7
6
6
7
7
6 68:1811 j46:9144 7
6
0
7
6
6
7
6 170:8669 j105:1733 7
6 0:0005 j0:0003 7
7
6
6
7
6 239:0480 j152:08774 7
6 0:0013 j0:0010 7
7
6
7; w

7 6
v
6 0:1000 j2:5611 7 7 6 0:0005 j0:0001 7

6
6
7
7
6
6 0:0039 j0:0037 7
7
2:449 j6:2889
6
6
7
7
6
6 0:0042 j0:0188 7
7
3:5529 j8:2378
6
7
7
6
6 0:0017 j0:0447 7
7
6
0:0009 j0:0057
6
7
7
6
4 0:1917 j0:1403 5
4 0:0042 j0:0224 5
0:0067 j0:0301
0:9683
2
p71 0:1427; p72 0:3021; p73 0:0618:
7
Eq. (5.63) as r7 4:8776, which is greater than 1. Hence, k
1:5051 j9:1860 is an electromechanical oscillation mode.
9 0:1736 j7:0590
Similarly, right and left eigenvectors corresponding to k
are calculated as
2
2
3
3
0:0830 j2:4518
0:0018 j0:0581
6 0:0192 j1:8299 7
6 0:0085 j0:1649 7
6
6
7
7
6 0:0638 j0:62186 7
6 0:0082 j0:0985 7
6
6
7
7
6 108:9599 j6:3716 7
6 0:0013 j0:0001 7
6
6
7
7
6 81:3708 j2:8540 7
6 0:0037 j0:0003 7
6
6
7
7
6 27:5890 j3:5176 7
6 0:0022 j0:0002 7
6
7; w
7

9 6
v
6 0:6337 j0:5454 7 9 6 0:0015 j0:0016 7

6
6
7
7
6 0:9340 j0:2801 7
6 0:0066 j0:0195 7
6
7
7
6
6 0:0030 j0:0134 7
6 0:2936 j0:0842 7
6
7
7
6
6 0:0973 j0:0918 7
6 0:0012 j0:0014 7
7
7
6
6
5
4 0:0029 j0:0014 5
4
0:8031
0:0009 j0:0004
0:5441 j0:0031i
223
p91 0:0031; p92 0:1254; p93 0:4717:
9 0:1736
Eq. (5.63) as q9 34:3291, which is greater than 1. Hence, k
j7:0590 is also an electromechanical oscillation mode.
According to Eq. (5.66), mode shape of electromechanical oscillation mode can
be identied from the elements of right eigenvector corresponding to
Dxj ; j 1; 2; 3, which is the 4th, 5th, and 6th element of right eigenvector,
7 1:5051 j9:1860, the elements of right eigenvector are
respectively. For k

v74 0:0000 j0:0000 0\42:2017
75 0:0005 j0:0003 0:0006\29:5912
v
76 0:0013 j0:0010 0:0017\144:0796
v
The mode shape is illustrated in Fig. 5.8. It shows that the electromechanical
oscillation associated with this mode is that of G3 against G2.
-4
x 10
G2
-2
-4
-6
-8
G3
-10
-12
-10
-8
-6
-4
-2
x 10
7 1:5051 j9:1860
Fig. 5.8 Mode shape of k
-4
224
x 10 -3
1
G1
0
G2
G3
-1
-2
-3
-4
-4
-3
-2
-1
4
-3
x 10
9 0:1736 j7:0590
Fig. 5.9 Mode shape of k
9 0:1736 j7.0590, the elements of right eigenvector are

Similarly, for k

v94 0:0013 j0:0001 0:0013\3:2273

v95 0:0037 j0:0003 0:0037\175:6366
96 0:0022 j0:0002 0:0022\173:8540
v
The mode shape is illustrated in Fig. 5.9. It indicates that the electromechanical
oscillation associated with this mode is that of G1 against G2 and G3.
5.3.2.2
Selective Reduced-Order Calculation of Electromechanical

Oscillation Modes of Example Power System
From the HeffronPhillips model of example power system, the compact model of
system of Eq. (5.68) in the form of transfer function matrix is obtained as
Fs) s2 M sD x0 K1 K2 I sTA K3 sT0d0 KA K6 ]1
I sTA K4 KA Kk ]
225
The Hooke-Jeeves optimum searching method in nonlinear programming can be

used to nd the solution of following objective function of Eq. (5.74) to calculate
electromechanical oscillation modes as follows.

DEL[Fk

f(k
5:74
p
Step 1: Calculate eigenvalues of matrix x0 M1 K1 as gi ; i 1; 2. Thus, gi 6
0; j 1; 2 is the angular natural oscillation frequency of the electromechanical
i 0 0 jp
gi on the complex plane as the starting
oscillation modes. Choose k
point of optimum searching.
Step 2: Choose an initial searching scale S0 [ 0, a minimum searching scale
i k)
SF [ 0, and an accelerating factor a [ 0. Search on the complex plane from k

i k)
to ki k 1. Three different patterns of searching on the complex plane from k

to ki k 1 are illustrated in Fig. 5.10 where S is the searching step and a 2. In
S\fk
i or fk
jS\fk
i , the searching is successful in
Fig. 5.10 if fk
i
i
horizontal and vertical direction. Otherwise, it is a failure. From Fig. 5.10, it can be
seen that (1) pattern 1there is one successful searching in both horizontal and
pattern 1
pattern 2
pattern 3
complex plane
the starting point i(k)
S
the ending point i(k+1)
searching scale
the point where the searching is successful(S reduces)

the point where the searching fails(S does not reduce)
the serach is accelerated along the successful direction
Fig. 5.10 Illustration of searching patterns of the Hooke-Jeeves optimum searching method in one
step
226
vertical direction; (2) pattern 2only one searching in the horizontal direction is
successful; (3) pattern 3only one searching in the vertical direction is successful.
Step 3: If searching in Step 2 fails, reduce the searching scale S by half and then go
back to Step 2;
i (kF ), the searching scale becomes less than SF , stop the searching and
Step 4: If at k

take ki (kF ) as the solution of objective function of Eq. (5.74), i.e., the ith oscillation
mode of the power system.
Figure 5.11 presents the flow chart of the Hooke-Jeeves optimum searching
method as explained above for the selective reduced-order calculation of the
electromechanical oscillation modes
i (kF ) to the
The minimum searching scale SF [ 0 is the maximum distance of k

position where the real solution ki locates as shown in Fig. 5.12, where the last
searching scale is not greater than 2SF . Therefore, SF denes the accuracy of
optimum searching which is fully controllable. Of course, theoretically, the algorithm can reach a solution which is as accurate as it is wished.
In the case that two or more starting points might converge to the same solution,
preventing techniques, such as the addition of a penalty or barrier function into the
objective function [3], can be applied in the direct searching.
The computational complexity (CC) [3] or computational cost (CC) [4] normally
is used to estimate the computational efciency of an algorithm where one operation of an addition or multiplication is dened as one CC. The CC of multiplication
of two n-order full matrices (none of them is a diagonal matrix) is about n3 and the
3
inverse calculation of an n-order full matrix is n3 . Thus, the CC to form the
polynomial matrix model of Eq. (5.74) (i.e. forming all coefcient matrices in F(s))
is estimated as 2n3 =3 32n3 . The CC in calculating the determinant of an n-order
complex matrix is about 4n3 =3. Thus, the CC of the Hooke-Jeeves optimum
searching in one step is about 16n3 =3. Therefore, the CC of the Hooke-Jeeves
optimum searching method is (kHS KHF )16n3 =3, where kHS is the times of successful searching and KHF the times of searching failure. KHF can be estimated as
kHF logS0 =SF = log R, where R is the ratio of reducing the searching scale when
the searching in one step fails. If R = 2 S0 0:01; SF 0:000001, we can have
kHF 13. Hence, the total CC of the reduced-order method is about
102n3 kHS 16n3 =3 on2 .
Figures 5.13 and 5.14 show the trajectories of optimal searching starting from (0,
7.5128) and (0, 10.5647), respectively. Searching results of the electromechanical
1 0:1736 j7:0590 and k
2 1:5051 j9:1860,
oscillation modes are k
which are exactly as same as those obtained in the previous section where they are
9 0:1736 j7:0590 and k
7 1:5051 j9:1860.
denoted as k
227
Start
Step 1
Choose i (0)=0+j i ,i=1,2 as the searching

start points,where i are the eigenvalues of
0.
matrix - 0 M -1K 1 and
i
Step 2
Initalization:i=k=m=1, d1 1 , d 2 j ,
set the value of S0 , SF and .
No
No
Yes
f( i (k)-d mS)<f( i (k))?
(k+1)= i (k)
(k+1)= i (k)+d mS
(k+1)= i (k)-d mS
No
No
Yes
f( i (k)+d mS)<f( i (k))?
Yes
m<2?
m=m+1
Yes
f( i (k+1))<f( i (k))?
i
(k+1)= i (k)+a( i (k+1)- i (k)),
k=k+1,m=1
Step 3
S= S 2,
i
No
S<SF?
(k+1)= i (k),
Yes
k=k+1,m=1
Step 4
i
(k) is the solution
End
Fig. 5.11 Flow chart of Hooke-Jeeves optimum searching method
228
SF
*
i
(k F )
point at each step
contour of S
searching succeeds
searching fails
Fig. 5.12 Illustration on the searching to arrive at its nal point
7.6
Imaginary part
48000000
of eigenvalue
7.5
Start Point
40000000
7.4
32000000
24000000
7.3
7.2
1600
0000
8000
000
7.1
Final Point
6.9
Real part of
eigenvalue
6.8
-0.25
-0.2
-0.15
-0.1
-0.05
1 calculation using direct searching algorithm

Fig. 5.13 Path of k
0.05

11
229
Imaginary part
of eigenvalue
10.8
900000000
10.6
Start Point
70000000
10.4
500000
000
10.2
10
3000
9.8
9.6
0000
100000
000
9.4
50000
000
9.2
Real part
of eigenvalue
Final Point
9
-1.6
-1.4
-1.2
-1
-0.8
-0.6
-0.4
-0.2
2 calculation using direct searching algorithm

Fig. 5.14 Path of k
5.3.2.3
Coordinated Design of PSSs
1 is 0.0246 and that of k

2 is 0.1617. Now PSSs will be installed
Damping ratio of k
1 to 0.1. Parameters of PSSs can
in three machines to enhance the damping ratio of k
1 to
be set by use of the method introduced previously in Sect. 5.2.2 to move k

k1 0:7094 j7:0590. Hence, following objective function is set up

2

2
1 p Re k
p Im k
1 p Im k
p
f p Re k
1
1
where p Dpss1 Dpss2 Dpss3 and Dpssi is the coefcient of damping torque
provided by the PSS installed in the ith generator.
Choose p0 0 0 0 with f p0 0:2872 to start the optimal searching of
the solution of the above objective function by use of the method introduced in
Sect. 5.2.2.2. In 52 iterations, the searching stops to have p52
1 is moved to k
c1 .
21:73 11:84 21:94 with f p52 9:2687 1015 , k
Figure 5.15 is the trajectories of Dpssi ; i 1; 2; 3 in respective to the iterative
searching. The optimal searching method used is the Hooke-Jeeves optimum
searching. Figure 5.16 is the trajectory of the objective function during the optimum
1 .
searching with respect to the movement of oscillation mode k
230
25
Designed damping
20
15
10
5
Dpss1
Dpss2
Dpss3
10
20
30
40
50
60
Iterations of direct search method
Fig. 5.15 Trajectories of Dpssi , i = 1, 2, 3 in respective to the iterative searching
Objective function
Start Point
Projection
Final Point
Imaginary part of eigenvalue
Real part of eigenvalue
1
Fig. 5.16 Trajectory of the objective function in respective to the movement of k

Table 5.6 Parameter of the
PSSs
G1
G2
G3
231
Kpss
T2
T4
4.1272
1.4864
3.0945
0.096
0.095
0.1
0.096
0.095
0.1
From Fig. 5.5, the forward path matrix from the stabilizing signals of PSSs to the
electromechanical oscillation loops of generators is obtained as

K2 KA K k
T0 I k
TA K6 KA 1
Fpss k
1
1 d0
1
3
3
2
4:0590\29:34 1:7719\153:70 1:0376\155:74
7
6
4 0:6433\113:76 6:1802\28:79 2:6963\136:09 5
1:1801\125:70 2:6184\140:40 5:3227\31:52
The forward path of the PSS installed in each generator as the diagonal element
of the above matrix is
4:0590\29:34
Fpss11 k
1
6:1802\28:79
Fpss22 k
1
5:3227\31:52
Fpss33 k
1
Let the transfer function of the PSS be
Tpssi s Kpssi
1 sT2i 1 sT4i
; T1i T3i 0:05
1 sT1i 1 sT3i
By use of the phase compensation method for the PSS to compensate the for ; i 1; 2; 3 to provide a pure damping torque
ward path of the ith generator Fii k

i

Dpssi , where Dpss1 Dpss2 Dpss3 p52 21:73 11:84 21:94 , parameters
of PSSs are set and listed in Table 5.6.
With PSSs installed, the closed-loop state equation is
sDX Apss DX
where

Apss
Apss11
Apss21
Apss12
Apss22
Apss13
Apss23
232
Apss11
0
6
0
6
6
0
6
6
6 0:0495
6
6
6 0:1195
6 0:1761
6
6
6 0:0031
6
4 0:2004
0:2172
0
0
0
0:0310
0:1731
0:1362
0:0001
0:3121
0:1717
0
0
0
0:0185
0:0537
0:3123
0:0030
0:1117
0:3889
Apss12
0
0
6
0
0
6
6
0
0
6
6
6 0:0521 0:0191
6
6
6 0:0801 0:2396
6 0:0614 0:1073
6
6
6 0:1327 0:0119
6
4 0:0202 0:5261
0:0768
0:2119
2
Apss13
Apss21
0
60
6
60
6
6
60
6
6
60
60
6
6
60
6
40
0
67:5657
75:7691
6 595:9666 314:8297
6
6 779:5679 498:8968
6
6
6 0:0951
0:0596
6
6
0:7703
0:4825
6
6 0:2270
0:3289
6
6
6 0:8193
1:1874
6
4 0:3521
0:2724
3:9227
3:0351
314:1592
0
0
314:1592
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0:0119
0:0317
0:3860
0:0111
0:1277
0:6306
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0:1116
0
0
0
0
0
0
0
0
0
0
0
143:3348
281:1369
1278:4646
0:0355
0:2878
0:1020
0:3681
0:6246
6:9579
0
0
0
0
0
0
0
0:1667
0
3
0
7
0
7
314:1592 7
7
7
7
0
7
7
0
7
7
0
7
7
7
0
7
5
0
0
3
0
0 7
7
0 7
7
7
0 7
7
0 7
7
0 7
7
7
0 7
7
0 5
0:1698
3
0
07
7
07
7
7
07
7
07
7
07
7
7
07
7
05
0
0
0
0
20
162:048
0
0
0
0
3
0
0
0
0 7
7
0
0 7
7
7
0
0 7
7
0
0 7
7
20
0 7
7
7
72:2
0 7
7
0
20 5
0
222:8
Apss22
8731:9338 710:95188
6 4315:4400 5291:9761
6
6 4691:1745 1828:9008
6
6
6 0:1000
0:0367
6
6
6 0:8099
0:2977
6 0:1521
0:4553
6
6
6 0:5491
1:6435
6
4 0:1227
0:2146
1:3673
2:3902
2
Apss23
0
6
0
6
6
0
6
6
6 20
6
6
6 77:648
6
0
6
6
6
0
6
4
0
0
10000
0
0
0
20
0
0
0
0
233
678:1403
1729:6500
4789:5838
0:0229
0:1855
0:0603
0:2176
0:7719
8:5990
0
0
0
0
0
20
34:2
0
0
0
10000
0
0
0
0
20
0
0
50
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
20
111:4
0
50
0
0
0
0
0
0
0
3
0
0 7
7
50 7
7
7
0 7
7
0 7
7
0 7
7
7
0 7
7
0 5
0
3
0
0 7
7
10000 7
7
7
0 7
7
0 7
7
0 7
7
7
0 7
7
0 5
20
Eigenvalues of state matrix are calculated as

1 44:7474 j0; k
2;3 23:7299 j27:9946; k
4 33:1534 j0,
k
5;6 19:9770 j11:9203; k
7;8 12:9863 j13:6705;
k
9;10 0:6094 j7:0235; k
11;12 4:5207 j8:0856;
k
13 0:2772 j0,k
14 0 j0, k
15 23:6389 j0,
k
17 17:7485 j0,k
18 15:9316 j0
16 12:1456 j0,k
k
c1 0:6094 j7:0235, slightly away from the
1 is moved to k
Obviously, k

0:7094 j7:0590. In Sect. 6.4.1, it is examined why
target required position k
1
this happens.
Simulation results of example power system without and with PSSs installed are
presented in Fig. 5.17. At 0.5 s of simulation, a 3-phase short-circuit occurs at Bus
1, and 0.1 s later the short-circuit fault is cleared. Fig. 5.17, conrms the effectiveness of PSSs designed by use of the introduced method.
234
2- 1(deg)
With PSS
Without PSS
time(ms)
Fig. 5.17 Simulation results with a 3-phase short-circuit fault at Bus 1
References
1. Anderson PM, Fouad AA (1977) Power system control and stability. The Iowa State University
Press
2. Wang XF, Song YH, Irving M (2011) Modern power systems analysis. Springer, Berlin
3. Bazarra MS et al (1993) Non-linear programming. Wiley, Hoboken
4. Marchuk GI (1994) Numerical methods and applications. CRC Press Inc, Boca Raton
Chapter 6
Multi-machine Power System Installed

with Thyristor-Based FACTS Stabilizers
6.1

System Installed with Thyristor-Based FACTS
Stabilizers
6.1.1
HeffronPhillips Model [1]
6.1.1.1

Installed with an SVC Stabilizer
Figure 6.1 shows an N-machine power system with nodes of internal voltage of
generators added. From Fig. 6.1, the following network equations can be
established.
0
Ig
Y11
Y21
Y12
Y22
Vm
Eg

6:1

T

g E
m is the
g2 E
gN T ; V
g1 E
where Ig Ig1 Ig2 IgN ; E
g
vector of voltage at all non-internal nodes of generators in the power system and E

is dened by Eq. (5.34). Deleting all non-internal nodes is to remove Vm in the
network equations of Eq. (6.1) to have
Ig Y
21 Y
22 E
g
1 Y
12 Y
11
6:2
where only N internal nodes of generators are kept. Comparing Eqs. (6.2) and (5.34),
it can be seen that in fact
Y
21 Y
1 Y
12 Y
22
Y
11

DOI 10.1007/978-1-4899-7696-3_6
6:3
235
236
Fig. 6.1 Network of an

N-machine power system
including internal nodes of
generators
E g1
Eg 2
E gN
Ig1
Ig 2
IgN
x d1 '
xd2 '
x dN '
V g1
V g2
V gN
Transmission network
Fig. 6.2 An N-machine

power system installed with
an SVC-based stabilizer
E1
E2
Ig1
YN
EN
Ig 2
IgN
X d1 '
Xd 2 '
X dN '
V g1
Vg 2
V gN
YN
Node 1
y11
jbsvc
Now consider an N-machine power system installed with an SVC stabilizer as

shown in Fig. 6.2. Without loss of generality of discussion, it can be assumed that
(1) the SVC stabilizer is installed at node 1 in the transmission network, and
(2) after other nodes are deleted in the network equations of the N-machine power
system, there are only N internal nodes of generators and node 1 left. Hence, the
network equations with those N + 1 nodes left are
0
Ig
11
Y
21
Y
12
Y
22
Y
1n
V
g
E
11 y11 jbsvc and jbsvc is the equivalent admittance of the SVC.

where Y
6:4
6.1 Mathematical Model of a Multi-Machine Power System
237

1
21 Y
12 E
g Y
g
22 E
Y
y11 jbsvc

1
g Y
21 Y
12 Y
22 E
E
g
Y
y11 jbsvc
Ig
6:5

From Eq. (6.5), it can be seen that the elements of network admittance matrix, Y,
can be denoted as
yij y0ij yij bsvc
6:6

Igi idi jiqi
N
X
yij E0qj ej90
dj di
aij dj di
xqj x0dj iqj ejaij dj di
j1
N
X
y0ij E0qj ej90
6:7
j1
N
X
yij bsvc E0qj ej90
dj di
j1
0ij y0ij ejaij . From Eq. (6.7), it can be obtained that

where y
idi
N
X
y0ij E0qj sinaij dij xqj x0dj iqj cosaij dij
j1
N
X
Reyij bsvc E0qj sin dij xqj x0dj iqj cos dij
j1
N
X
Imyij bsvc E0qj cos dij xqj x0dj iqj sin dij
j1
iqi
N
X
6:8
yij E0qj cosaij dij xqj x0dj iqj sinaij dij
j1
N
X
Reyij bsvc E0qj cos dij xqj x0dj iqj sin dij
j1
N
X
Im[yij bsvc E0qj sin dij xqj x0dj iqj cos dij
j1
where Re
yij bsvc and Imyij bsvc denote the real and imaginary parts of yij bsvc ,
respectively.
238

DId Fdd Dd Gdd DE0q Hdd DIq Ldd Dbsvc
DIq Fqq Dd Gqq DE0q Hqq DIq Lqq Dbsvc
6:9

DId Fd Dd Gd DE0q Ld Dbsvc
DIq Fq Dd Gq DE0q Lq Dbsvc
6:10
Substituting Eq. (6.10) into Eqs. (5.40) and (5.41), the HeffronPhillips model of
the N-machine power system installed with the SVC stabilizer can be obtained as
Dd_ xo Dx
Dx_ M1 K1 Dd K2 DE0q DDx KP Dbsvc
0
0
0
DE_ q T01
d0 K3 DEq K4 Dd DEfd Kq Dbsvc
6:11
0
1
0
DE_ fd T1
A DEfd TA KA K5 Dd K6 DEq KV Dbsvc
6.1.1.2

Installed with a TCSC or TCPS Stabilizer
Without loss of generality of discussion, it can be assumed that a TCSC stabilizer is

installed between nodes 1 and 2 in an N-machine power system as shown in
Fig. 6.3. The following network admittance matrix can be formed with only the
internal nodes of generators, nodes 1 and 2 left,
0
Ig

Y
11
Y21
12
Y
22
Y
12
V
g
E

6:12
where
2
11
Y
12
V
y11 1
z12 jxt csc
4
y21 1
z12 jxt csc

1 V
2 T
V
y12 z
3
jxt csc 5
1
12
y22 z 1jx
12
t csc
Hence, the network admittance matrix with only the internal nodes of generators
left is

TCSC-based stabilizer
E g1
239
Eg 2
E gN
Ig1
Ig 2
IgN
x d1 '
xd2 '
x dN '
V g1
Vg 2
V gN
YN
Y
y12 (y 21 )
V1
V2
Node 1
z12
Node 2
x t csc
y11
y 22
Y
21 Y
1 Y

Y
11 12 Y22
6:13
Obviously, the element of the network admittance matrix yij can be denoted as
yij y0ij yij xt csc
6:14

Igi idi jiqi
N
X
yij E0qj ej90
dj di
aij dj di
j1
N
X
y0ij E0qj ej90
6:15
j1
N
X
yij xt csc E0qj ej90
dj di
j1
Similar to Eqs. (6.7)(6.10), by using Eq. (6.15), the HeffronPhillips model of

the N-machine power system installed with the TCSC stabilizer can be obtained as
Dd_ xo Dx
Dx_ M1 K1 Dd K2 DE0q DDx KP Dxt csc
0
0
0
DE_ q T01
d0 K3 DEq K4 Dd DEfd Kq Dxt csc
0
0
1
0
DE_ fd T1
A DEfd TA KA K5 Dd K6 DEq KV Dxt csc
6:16
240

TCPS-based stabilizer

Eg 2
E g1
E gN
Ig1
Ig 2
IgN
x d1 '
xd2 '
x dN '
V g1
Vg 2
V gN
YN
y12 (y 21 )
V1
V2
Node 1
Node 2
z12
k :1
y11
y 22
Similarly, for the N-machine power system with a TCPS stabilizer installed
between nodes 1 and 2 as shown in Fig. 6.4, where k kej/ , it can have
0
Ig
12
Y
22
Y
11
Y
21
Y
12
V
g
E

6:17
where
2
11 4
Y
12n
V
y11
1
z 12
k
j/
y21 e
kz12

1 V
2 T
V
j/
y12 e
kz
12
y22 z112
3
5
Hence, the network admittance matrix with only the internal nodes of generators
left can be obtained as
Y
21 Y
1 Y

Y
11 12 Y22
6:18
and the element of the network admittance matrix yij can be denoted as
yij y0ij yij /
6:19
241
Hence, for the system with the TCPS stabilizer installed, it can have
Igi idi jiqi
N
X
yij E0qj ej90
dj di
aij dj di
j1
N
X
y0ij E0qj ej90
6:20
j1
N
X
yij /E0qj ej90
dj di
j1
By taking the similar procedure from Eqs. (6.7) to (6.10), from Eq. (6.20), the
HeffronPhillips model of the N-machine power system installed with the TCPS
stabilizer can be obtained as
Dd_ xo Dx
Dx_ M1 K1 Dd K2 DE0q DDx KP D/
6:21
0
0
DE_ q T01
d0 K3 DEq K4 Dd DEfd Kq D/
0
0
1
0
DE_ fd T1
A DEfd TA KA K5 Dd K6 DEq KV D/
The HeffronPhillips model of the N-machine power system installed with an

SVC stabilizer, TCSC stabilizer, or TCPS stabilizer can be expressed in a unied
form as shown in Fig. 6.5. In Fig. 6.5, Dufactss Dbsvc for the SVC stabilizer,
Dufactss Dxt csc for the TCSC stabilizer, and Dufactss D/ for the TCPS stabilizer, respectively.
K1
0 I
s
(sM + D) 1
u facts s
KV
Kq
KP
K4
K5
K2
+
1
(K 3 + sTd0 )
(I + sTA ) 1 K A
u pss
K6
Fig. 6.5 Unied HeffronPhillips model of an N-machine power system installed with a
thyristor-based FACTS stabilizer
242
6.1.2
General Linearized Model of an N-Machine Power

System Installed with Multiple Thyristor-Based
FACTS Stabilizers
In this section, a more general procedure to establish the linearized model of an

N-machine power system installed with multiple thyristor-based FACTS stabilizers
is introduced. In the established model, the generator can be described by either the
reduced-order dynamic model which is used in the above section to establish the
HeffronPhillips model, or the full linearized model given by Eqs. (5.44)(5.47). In
addition, the normal control functions of thyristor-based FACTS controllers and the
full dynamics of stabilizers are also included in the established model.
6.1.2.1
Linearized Model of Generators
When full linearized mathematical model of generators is used, by substituting

Eq. (5.51) into Eq. (5.48), it can be obtained that
X_ g Agxy Xg Bg Dupss Bgxy DVxy
DIxy Cgxy Xg Dgxy DVxy
6:22
where
Agxy Ag Bgv BgV
Bgxy Bgv Tg0
Cgxy T1
g0 Cg BgI
Dgxy 0
In the case that the reduced-order model of generators which is used for
establishing the HeffronPhillips mode, from Eq. (2.31), it can have
iqi
vgdi
E0 vgqi
; idi d 0
xqi
xd
6:23
Linearization of the above equation is

0
Diqi
DEq Dvgqi
Dvgdi
; Didi
xqi
x0q
6:24
243
From Eq. (6.24) and (5.39), it can be obtained that

DIdq E0 DE0iq Xdq1 DVdq
DP Vdq0 DIdq Idq0 DVdq
DEq DE0q X0dd DIdq
6:25
DV V0dq DVdq
where
E0 feij g; eij
1
;
x0di
for i 1; 3; . . .; 2N 1; j 1; 2; . . .; N; otherwise eij 0

1
Xdq1 diagxdqi ; xdqi 0 ; for i 1; 3; . . .; 2N 1;
xdi
1
0; for i 2; 4; . . .2N
xdqi
xqi
Vdq0 fVdq0ij g; Vdq0ij vgdi0 ; Vdq0ij 1 vgqi0 ;
for i 1; 2; . . .; N; j 1; 3; . . .; 2N 1; otherwise Vdq0ij 0
Idq0 fIdq0ij g; Idq0ij igdi0 ; Idq0ij 1 igqi0 ;
for i 1; 2; . . .; N; j 1; 3; . . .; 2N 1; otherwise Idq0ij 0
X0dd fx0ddij g; x0ddij xdi x0di ;
for i 1; 2; . . .; N; j 1; 3; . . .; 2N 1; otherwise x0ddij 0
vgdi0
vgqi0
V0dq fV0dqij g; V0dqij
; V0dqij 1
;
Vgi0
Vgi0
for i 1; 2; . . .; N; j 1; 3; . . .; 2N 1; otherwise Vdq0ij 0
By substituting the above equation into (5.40), it can be obtained that
Dd_ xo Dx
Dx_ M1 DDx Vdq0 E0 DE0iq M1 Vdq0 Xdq1 Idq0 DVdq
0
0
0
0
0
01 0
DE_ q T01
d0 DEq Xdd E0 DEiq DEfd Td0 Xdd Xdq1 DVdq
0
DE_ fd
DIdq
0
1
1
T1
A DEfd TA KA V0dq DVdq TA KA Dupss
E0 DE0iq Xdq1 DVdq
6:26
244
which can be written as

2
3 2
32
3
Dd_
0
0
0
xo I
Dd
6
7 6
6
7
6 Dx_ 7 6 0 M1 D
M1 Vdq0 E0
0 7
76 Dx 7
6
76
0
6
7
7
0
6 DE_ 7 4 0
54 DE0q 5
0
T1
T1
q 5
4
d0 I Xdd E0
d0
0
0
0
0
T1
DE0fd
DE_ fd
A
3
3
2
2
0
0
6 M1 V X I 7
6 0 7
dq0 dq1
dq0 7
7
6
6
7DVdq 6
7Dupss
6
0
5
5
4
4
T01
X
X
0
d0
dd dq1
T1
A KA V0dq
DIdq 0
E0
Dd
6 Dx
6
0 6
4 DE0q
6:27
T1
A KA
7
7
7 Xdq1 DVdq
5
DE0fd
Obviously, the above equation can be written in the similar form to Eq. (5.48) as
X_ g Ag Xg Bg Dupss Bgv DVdq
DIdq Cg Xg Dg DVdq
6:28
where
3
2
0
Dd
0
xo I
1
6 Dx 7
6 0 M1 D
M Vdq0 E0
7
6
6
Xg 6
7; Ag 6
0
4 DE0q 5
40
0
T1
d0 I Xdd E0
DE0
0
0
0
3
3
2 fd
2
0
0
7
7
6
6
6 M1 Vdq0 Xdq1 Idq0 7
6 0 7
7
7
6
6
Bgv 6
7; Bg 6
7;
01 0
7
7
6
6
T
X
X
0
dq1
d0
dd
5
5
4
4
2
T1
A KA V0dq
Cg 0
E0
0
0
T1
d0
3
7
7
7;
5
T1
A
T1
A KA
0 ; Dg Xdq1
By substituting Eq. (5.51) into the above equation, it can be obtained that
X_ g Agxy Xg Bg Dupss Bgxy DVxy
DIxy Cgxy Xg Dgxy DVxy
6:29
245
where
Agxy Ag Bgv BgV
Bgxy Bgv Tg0
Cgxy T1
g0 Cg BgI
Dgxy T1
g0 Dg
The format of Eqs. (6.22) and (6.29) is exactly the same, which will be used as
the general linearized model of generators in the N-machine power system with no
PSSs installed.
Let the transfer function of the ith PSS be Gi s and the feedback signal Dxi , that is
Dupssi Gi sDxi
6:30
If the state-space realization of Eq. (6.30) is

X_ pssi Apssi Xpssi Bpssi Dxi
6:31
Dupssi Cpssi Xpssi Dpssi Dxi
State-space representation of L PSSs installed in the N-machine power system

can be written as
X_ pss Apss Xpss Bpss Dx
6:32
Dupss Cpss Xpss Dpss Dx

where
3
2
Xpss1
Bpss1
6 Xpss2 7
6 0
7
6
6
7
6
6
6 .. 7; Bpss 6 ..
4 . 5
4 .
2
Xpss
Apss
Dpss
XpssL
2
0
Apss1
6 0
Apss2
6
6
..
6 ..
4 .
.
0
0
2
0
Dpss1
6 0
Dpss2
6
6
..
6 ..
4 .
.
0
0
0
...
...
...
...
...
...
...
...
0
0
..
.
0
Bpss2
..
.
0
3
...
...
...
...
7
7
7; Cpss
7
5
ApssL
3
0
0 7
7
7
.. 7
. 5
DpssL
0
0
..
.
BpssL
Cpss1
6 0
6
6
6 ..
4 .
0
2
3
7
7
7;
7
5
0
Cpss2
..
.
0
...
...
...
...
0
0
..
.
CpssL
3
7
7
7;
7
5
246
By substituting Eq. (6.32) into Eqs. (6.22) or (6.29), it can be obtained that
X_ gp Agp Xgp Bgp DVxy
6:33
DIxy Cgp Xgp Dgp DVxy

where
Xg
Xgp
Agxy Bg DDpss
Bg Cpss

;
; Agp
Xpss
BDpss
Apss

Bgxy
Dgp
; Cgp Cgxy 0 ; Dgp

;
0
0
2
2
3
3
0
0
0
0
0
0
6
6
7
7
4 0 Dpss 0 5; BDpss 4 0 Bpss 0 5
Bgp
DDpss
Equation (6.33) is the linearized model of generators with the PSSs installed.
6.1.2.2
Linearized Model of an SVC Installed in the N-Machine Power

System
For a thyristor-controlled reactor and xed capacitor (TCR-FC) type of SVC with
an SVC stabilizer as shown in Fig. 3.1, from Eq. (3.1), it can be obtained that for
the ith SVC installed in the N-machine power system
bsvci
1
2ai sin 2ai

xsvcci
2pxsvcli
6:34
Without loss of generality, denote the installing location of the SVC as the ith
node in the power system. It can be assumed that the transfer function of voltage
controller and stabilizer in Fig. 3.1 is Tvsvci s and Tssvci s, respectively, and the
feedback signal of the SVC stabilizer is the integral signal of the active power, Pij ,
delivered through the node where the SVC is installed as shown in Fig. 6.6. From
Figs. 3.1 and 6.6, it can have
Fig. 6.6 An SVC installed on

the ith node in the N-machine
power system
Vi
zij
Pij
Ii
jbsvci
Vj
Iij
247
ai avi asi
6:35
avi Tvsvci sVi Viref

1
asi Tssvci s Pij Pijref
s
6:36
where
where Viref and Pijref are the reference signal of the SVC voltage and stabilizing
control, respectively.
Linearization of Eqs. (6.34) and (6.36) is
Dbsvci
1 cos 2ai0
Davi Dasi
pxsvcli
6:37
Davi Tvsvci sDVi

1
Dasi Tssvci s DPij
s
where
DVi
h
viy0
vix0
Dvix
Dviy vVix0
i0
Vi0
Vi0
viy0
Vi0
6:38
i Dv
ix
Dviy
h
vVix0
i0
i
viy0
Vi0 DVixy

Pij Re
i V
j
V
Regij jbij V
V

iV
jV
V
i
i
i
zij
6:39
gij v2xi v2yi bij vxi vyi vyi vxj

where * denotes the conjugate of a phasor. Linearization of the above equation
gives
DPij 2gij vxi0 Dvxi vyi0 Dvyi bij vyj0 Dvxi vxj0 Dvyi
bij vxi0 Dvyj vyi0 Dvxj
2gij vxi0 bij vyj0 Dvxi 2gij vyi0 bij vxj0 Dvyi
bij vxi0 Dvyj vyj0 Dvxj

Dvxi
v
b
v
2g
v

b
v
2g
ij yj0
ij xj0
ij xi0
ij yi0
Dvyi

Dvxj
bij vxi0 bij vyi0
Dvyj

2gij vxi0 bij vyj0 2gij vyi0 bij vxj0 DVxyi
bij vxi0 bij vyi0 DVxyj
psvci DVxyi p0svcj DVxyj
6:40
248
Hence, from Eqs. (6.38) and (6.40), it can have

h
i
v
Davi Tvsvci s vVxi0 Vyi0 DVxyi
i0
i0
1
Dasi Tssvci s psvci DVxyi p0svcj DVxyj
s
6:41
Let the state-space realization of the voltage controller and stabilizer of the SVC
given by Eq. (6.41) be
X_ VSVCi AVSVCi XVSVCi BVSVCi DVxyi
6:42
Davi CVSVCi XVSVCi DVSVCi DVxyi

X_ SSVCi ASSVCi XSSVCi BSSVCi DVxyi BSSVCj DVxyj
Dasi CSSVCi XSSVCi DSSVCi DVxyi DSSVCj DVxyj
6:43
According to Fig. 6.6, the injected current from the SVC into the ith node is
Ii ixi jiyi jbsvci V
i jbsvci vxi jvyi
6:44

ixi bsvci vyi ; iyi bsvci vxi
6:45

Dixi
Diyi

Dvxi
0
bsvci0
vyi0
Dbsvci
0
bsvci0 Dvyi
vxi0

bsvci0
1 cos 2ai0 vyi0

Davi Dasi
pXsvcli
0
vxi0
0
bsvci0

DVxyi
6:46
When only the SVC voltage control function is considered, by substituting

Eq. (6.42) into (6.46), it can have
Dixi
Diyi

Dvxi
vyi0
bsvci0
0
Dbsvci
0
bsvci0 Dvyi
vxi0

1 cos 2ai0 vyi0
CVSVCi XVSVCi DVSVCi DVxyi DaSi

pxsvcli
vxi0

0
bsvci0
DVxyi
0
bsvci0
6:47
249
That is
DISVCi CSVCi XVSVCi DSVCi DVxyi bSSVCi Dasi
6:48
where

Dixi
1 cos 2ai0 vyi0
; CSVCi
CVSVCi ;
pxsvcli
Diyi
vxi0

1 cos 2ai0 vyi0
1 cos 2ai0 vyi0

DVSVCi ; bSSVCi
pxsvcli
pxsvcli
vxi0
vxi0
DISVCi
DSVCi
Equations (6.42) and (6.48) are the linearized model of the SVC without the
stabilizer. Feedback signal of the SVC stabilizer is DPij . Hence, output equation is
Eq. (6.40).
When the SVC stabilizer is considered, by substituting Eq. (6.43) into
Eq. (6.47), it can be obtained that
Dixi
Diyi

1 cos 2ai0 vyi0
CVSVCi XVSVCi CSSVCi XSSVCi
pxsvcli
vxi0

1 cos 2ai0 vyi0

DVSVCi DVxyi DSSVCi DVxyi DSSVCj DVxyj
pxsvcli
vxi0

bsvci0
0
DVxyi
0
bsvci0
6:49
Put the state equation of the SVC voltage controller of Eq. (6.42) and stabilizer
of Eq. (6.43) together
X_ SVCi ASVCi XSVCi BSVCi DVxyi BSVCj DVxyj
6:50
where

XSVCi

BSVCi
XVSVCi
XSSVCi

BVSVCi
BSSVCi

; ASVCi
; BSVCj
AVSVCi
0
0
0
ASSVCi

;
BSSVCj

DISVCi CSVCi XSVCi DSVCi DVxyi DSVCj DVxyj
6:51
250
where

DISVCi
Dixi
Diyi

CSVCi
CSSVCi
vxi0
vxi0

bsvci0
0
1 cos 2ai0 vyi0
DSVCi
DVSVCi DSSVCi
pxsvcli
0
bsvci0
vxi0

1 cos 2ai0 vyi0
DSVCj
DSSVCj
pxsvcli
vxi0
2ai0
1cos
pxsvcli
vyi0
2ai0
CVSVCi 1cos
pxsvcli
vyi0
Equations (6.50) and (6.51) are the linearized model of the SVC, including
voltage controller and stabilizer, installed at the ith node in the N-machine power
system.
6.1.2.3
Linearized Model of a TCSC Installed in the N-Machine Power

System
At the location in the N-machine power system where a TCSC is installed, two
extra nodes, denoted as i and j, at the terminals of the TCSC can be created as
shown in Fig. 6.7. From Fig. 6.7, it can be seen that the function of the TCSC is
electrically equivalent to the injection of current into two nodes,

Ii Vj Vi ; Ij Vi Vj
jxtcsci
jxtcsci
6:52
Vi
Pij
Vi
Vj
Ij
Ii
Vj
Ii
jx tcsci
Fig. 6.7 Creation of two extra nodes at the location where the TCSC is installed
Ij
251
That is
Ii ixi jiyi j vxj jvyj vxi jvyi
xtcsci
Ij ixj jiyj j vxi jvyi vxj jvyj
xtcsci
6:53

ixi
vyj vyi ; iyi
vxi vxj ;
xtcsci
xt csci
1
1
ixj
vyi vyj ; iyj
vxj vxi
xtcsci
xtcsci
6:54
The normal control function of the TCSC is the load flow regulation. Let the
transfer function of the TCSC load flow regulator be Tltcsci s. It can be assumed
that the transfer function of the TCSC stabilizer and feedback signal is Tstcsci s and
the integral of deviation of the active power delivered along the transmission line,
Pij , where the TCSC is installed. Control signal of the TCSC stabilizer can be
superimposed on the TCSC load flow controller, that is
1
xtcsci Tltcsci s Tstcsci sPij Pijref
s
6:55
where Pijref is the reference signal of the TCSC load flow and stabilizing control.
1
Dxtcsci Tltcsci s Tstcsci sDPij
s
6:56

Pij Re
i V
j
V
Re 1 V
V

iV
jV
V
i
i
i
jxtcsci
jxtcsci
1
xtcsci
6:57
vyi vxj vxi vyj

1
vyi0 vxj0 vxi0 vyj0 Dvtcsci
x2tcsci0
1
1
vxj0 Dvyi vyj0 Dvxi

vyj0 Dvxj vxi0 Dvyj
xtcsci0
xtcsci0
atcsc0 Dxtcsci atcsci DVxyi atcscj DVxyj
DPij
6:58
252
Fig. 6.8 TCSC with a PI load

flow controller superimposed
by a stabilizing signal

K ptcsc
Pij
+
K itcsc
s
x tcsci
x ltcsci
stcsci
where
1
vyi0 vxj0 vxi0 vyj0
x2tcsci0
h
i
1
1
vyj0 xtcsci0
vxj0
xtcsci0
h
i
1
1
vyj0 xtcsci0
vxi0
xtcsci0
atcsc0
atcsci
atcscj
For the case that the TCSC stabilizer is not installed, Eq. (6.56) becomes
Dxtcsci Tltcsci sDPij Dastcsci
6:59
where Dastcsci is the stabilizing signal. Without loss of generality, it can be assumed
that the TCSC load flow controller adopts a proportional and integral (PI) control
law as shown in Fig. 6.8. From Fig. 6.8, it can have
Dxtcsci Dastcsci Kptcsc DPij Dxltcsci
Dx_ ltcsci Kitcsc DPij
6:60

Dxtcsci
Dastcsci
Dxltcsci
1 Kptcsc atcsc0
1 Kptcsc atcsc0
Kptcsc
atcsci DVxyi atcscj DVxyj

1 Kptcsc atcsc0
6:61
Kitcsc atcsc0
Kitcsc atcsc0
Dxltcsci
Dastcsci
1 Kptcsc atcsc0
1 Kptcsc atcsc0
Kptcsc Kitcsc atcsc0
Kitcsc
1 Kptcsc atcsc0
6:62
Dx_ ltcsci

Dx_ ltcsci Altcsci Dxltcsci BLTCSCi DVxyi BLTCSCj DVxyj bltcsci Dastcsci
6:63
253

Dixi
That is
2
vyj0 vyi0 Dxtcsci
Dvyj Dvyi ;
xtcsci0
1
1
Diyi 2 vxi0 vxj0 Dxtcsci
Dvxi Dvxj ;
xtcsci0
xtcsci0
1
1
Dixj 2 vyi0 vyj0 Dxtcsci
Dvyi Dvyj ;
xtcsci0
xtcsci0
1
1
Diyj 2 vxj0 vxi0 Dxtcsci
Dvxj Dvxi
xtcsci0
xtcsci0
x2tcsci0
x21 vyj0 vyi0
6:64
tcsci0
Dixi
7
6
1
6 Di 7 6
vxi0 vxj0 7

2
7
x
6 yi 7 6
tcsci0
7Dxtcsci
76
6
7
1
4 Dixj 5 6
6 x2tcsci0 vyi0 vyj0 7
5
4
Diyj
x21 vxj0 vxi0
tcsci0
32
2
3
Dvxi
0 1 0
1
7
6
0 1 0 7
1 6
76 Dvyi 7
6 1
76
6
7
xtcsci0 4 0
1
0 1 54 Dvxj 5
Dvyj
1 0
1
0
6:65
By using Eq. (6.61), the above equation becomes

DITCSCij CTCSCij Dxltcsci bTCSCij Dastcsci DTCSCij DVxyij
where
DITCSCij
3
3
2
Dixi
Dvxi
6 Diyi 7
6 Dvyi 7
7
7
6
6
6
7; DVxyij 6
7;
4 Dixj 5
4 Dvxj 5
2
Diyj
bTCSCij
DTCSCij
6:66
Dvyj
x21 vyj0 vyi0
tcsci0
7
6
6 21 vxi0 vxj0 7
7
6 xtcsci0
7;
6
CTCSCij
7
1
1 Kptcsc atcsc0 6
6 x2tcsci0 vyi0 vyj0 7
5
4
x21 vxj0 vxi0
tcsci0
2
0 1

atcsci
0
1
0
Kptcsc
1 6
6
6
4
x
1 Kptcsc atcsc0
0
atcscj
0
1
tcsci0
1 0
3
0
1
1 0 7
7
7
0 1 5
1
0
6:67
254
Equations (6.63) and (6.66) are the linearized model of the TCSC without the TCSC
stabilizer installed. Output equation can be obtained from Eqs. (6.58) and (6.61) as
DPij dtcsci Dastcsci ctcsci Dxltcsci ptcsci DVxyi p0tcscj DVxyj
6:68
where
atcsc0
atcsc0
; ctcsci
;
1 Kptcsc atcsc0
1 Kptcsc atcsc 0

atcsc0 Kptcsc
atcsc0 Kptcsc
0
1
atcsci ; ptcscj 1
atcscj
1 Kpt csc atcsc0
1 Kptcsc atcsc0
dtcsci
ptcsci
When the TCSC stabilizer is installed as shown in Eq. (6.55), denote

Ttcsci s Tltcsci s s1 Tstcsci s. From Eqs. (6.56) and (6.58), it can have
Ttcsci s
1 atcsc0 Ttcsci s
T0t csc i satcsci DVxyi atcscj DVxyj
Dxtcsci
6:69
Let the state-space realization of Eq. (6.69) be

X_ TCSCi ATCSCi XTCSCi BTCSCi DVxyi BTCSCj DVxyj
6:70
Dxtcsci CTCSCi XTCSCi DTCSCi DVxyi DTCSCj DVxyj
6:71
By substituting Eq. (6.71) into Eq. (6.65), it can be obtained that

DITCSCij CTCSCij XTCSCi DTCSCij DVxyij
6:72
where
2
DITCSCij
DVxyij
x21 vyj0 vyi0
Dixi
7
6
6 21 vxi0 vxj0 7
6 Di 7
7
6
x
yi
7
6
tcsci0
7
6
7; CTCSCij 6
6 1 v v 7CTCSCi ;
4 Dixj 5
yj0 7
6 x2tcsci0 yi0
5
4
Diyj
x21 vxj0 vxi0
tcsci0
2
3
2
Dixi
0

6 1
6 Di 7
0
D
1
yi
TCSCi
6
7
6
6
6
7; DTCSCij
4 Dixj 5
xtcsci0 4 0
0
DTCSCj
Diyj
1
tcsci0
1
0
1
0
1
0
3
1
0 7
7
7
1 5
0
Equations (6.70) and (6.72) are the linearized model of the TCSC with stabilizer
installed between ith and jth nodes in the N-machine power system.
6.1.2.4
255
Linearized Model of a TCPS Installed in the N-Machine Power

System
Without loss of generality, it can be assumed that a TCPS is installed between

nodes i and j in an N-machine power system as shown in Fig. 6.9. From Fig. 6.9, it
can have
0 Iij
V
j/i
i
i I0 e
V
ij
6:73
0
j/i

I0 Vi Vj Vi e Vi Vi Vj
ij
zij
zij
j/i

i V

j V
j V
j ej/i
V
Iij Iij ej/i Vi Vj e
zij
zij
6:74
Hence, installation of the TCPS is electrically equivalent to the addition of a

current source at nodes i and j, respectively, as shown in Fig. 6.9, where
j/i
Ij Vi e Vi
zij
j/i

j
V
Ii Vj e
zij
6:75
Denote z1ij gij jbij . From Eq. (6.75), it can be obtained that
ixi gij vxj bij vyj cos/i 1 gij vyj bij vxj sin/i
iyi gij vxj bij vyj sin/i gij vyj bij vxj cos/i 1
ixj gij vxi bij vyi cos/i 1 gij vyi bij vxi sin/i
6:76
iyj gij vxi bij vyi sin/i gij vyi bij vxi cos/i 1
Vi
Pij
Vi
I ij
I ij '
k :1
k =1
z iijj
z iijj
Vj
Vj
Vi '
Ii
Fig. 6.9 Installation of a TCPS between node i and j in an N-machine power system
Ij
256

Dixi gij vxj0 bij vyj0 sin /i0 gij vyj0 bij vxj0 cos /i0 D/i
gij cos /i0 1 bij sin /i0 Dvxj bij cos /i0 1 gij sin /i0 Dvyj
Diyi gij vxj0 bij vyj0 cos /i0 gij vyj0 bij vxj0 sin /i0 D/i
gij sin /i0 bij cos /i0 1Dvxj bij sin /i0 gij cos /i0 1Dvyj
Dijx gij vxi0 bij vyi0 sin /i0 gij vyi0 bij vxi0 cos /i0 D/i
gij cos /i0 1 bij sin /i0 Dvxi bij cos /i0 1 gij sin /i0 Dvyi
Dijy gij vxi0 bij vyi0 cos /i0 gij vyi0 bij vxi0 sin /i0 D/i
gij sin /i0 bij cos /i0 1Dvxi bij sin /i0 gij cos /i0 1Dvyi
That is
DITCPSij bTCPSi D/i DTCPSi DVxyij
6:77
where
2
DITCPSij
2
3
3
Dixi
Dvxi
6 Di 7
6 Dv 7
6 yi 7
6 yi 7
6
7; DVxyij 6
7;
4 Dixj 5
4 Dvxj 5
Diyj
Dvyj
3
gij vxj0 bij vyj0 sin /i0 gij vyj0 bij vxj0 cos /i0
6 g vxj0 bij vyj0 cos / g vyj0 bij vxj0 sin / 7
i0
i0 7
ij
ij
6
6
7;
4 gij vxi0 bij vyi0 sin /i0 gij vyi0 bij vxi0 cos /i0 5
2
bTCPSi

DTCPSi
"
DiTCSCj
"
DiTCSCi
gij vxi0 bij vyi0 cos /i0 gij vyi0 bij vxi0 sin /i0

0
DTCSCij
DTCSCii
gij cos /i0 1 bij sin /i0

gij sin /i0 bij cos /i0 1
gij cos /i0 1 bij sin /i0
gij sin /i0 bij cos /i0 1
bij cos /i0 1 gij sin /i0
bij sin /i0 gij cos /i0 1

#
bij cos /i0 1 gij sin /i0
bij sin /i0 gij cos /i0 1
Typical application of a TCPS is to regulate power flow. Let the TCPS load flow
controller be implemented by a PI control law and a stabilizing control signal,
Dastcpsi , be added as shown in Fig. 6.10. It can have

Fig. 6.10 A PI TCPS load
flow controller superimposed
with a stabilizing signal
257
K ptcps
Pij
+
K itcps
s
x ltcpsi
+
stcpsi
D/i Dastcpsi Kpt c ps DPij Dxltcpsi

Dx_ ltcpsi Kit c ps DPij
6:78

j/i
i Iij V
Vi Vj e
Sij V
i
z
ij
2

j ej/i
gij jbij Vi Vi V
gij jbij V2i gij jbij vxi
6:79
jvyi vxj jvyj cos /i j sin /i
Hence,
Pij ReSij gij v2xi v2yi
gij vxi vxj vyi vyj cos /i vxi vyj vyi vxj sin /i
6:80
bij vxi vxj vyi vyj sin /i vxi vyj vyi vxj cos /i
DPij at c ps0 D/i atcpsi DVxyi atcpsj DVxyj
6:81
where
at c ps0 gij vxi0 vxj0 vyi0 vyj0 sin /i0 vxi0 vyj0 vyi0 vxj0 cos /i0
atcpsi
atcpsj
bij vxi0 vxj0 vyi0 vyj0 cos /i0 vxi0 vyj0 vyi0 vxj0 sin /i0
"
#T
2gij vxi0 gij vxj0 cos /i0 vyj0 sin /i0 bij vxj0 sin /i0 vyj0 cos /i0
2gij vyi0 gij vyj0 cos /i0 vxj0 sin /i0 bij vyj0 sin /i0 vxj0 cos /i0
"
#T
gij vxi0 cos /i0 vyi0 sin /i0 bij vxi0 sin /i0 vyi0 cos /i0
gij vyi0 cos /i0 vxi0 sin /i0 bij vyi0 sin /i0 vxi0 cos /i0
258

D/i
1
Dastcpsi
1 atcps0 Kptcps
Kptcps
1
atcpsi DVxyi atcpsj DVxyj

Dxltcpsi
1 atcps0 Kptcps
1 atcps0 Kptcps
6:82
Dx_ ltcpsi
atcps0 Kitcps
atcps0 Kitcps
Dxltcpsi
Dastcpsi
1 atcps0 Kptcps
1 atcps0 Kptcps
atcps0 Kitcps Kptcps
Kptcps
1 atcps0 Kitcps
6:83

Dx_ ltcpsi Altcpsi Dxltcpsi BLTCPSi DVxyi BLTCPSj DVxyj bltcpsi Dastcpsi
6:84

DITCPSij CTCPSij xltcpsi bTCPSij Dastcpsi DTCPSij DVxyij
6:85
where
1
bTCPSi
1 at c ps0 Kpt c ps

atcpsi
0
Kpt c ps
DTCPSi
1 at c ps0 Kpt c ps
0
atcpsj
bTCPSij CTCPSij
DTCPSij
6:86
Equations (6.84) and (6.85) are the linearized model of the TCPS without the
TCPS stabilizer installed. From Eqs. (6.81) and (6.82), the output equation can be
obtained as
DPij
at c ps0
at c ps0
Dastcpsi
Dxltcpsi
1 at c ps0 Kpt c ps
1 at c ps0 Kpt c ps
1

1 at c ps0 Kpt c ps
6:87
dtcpsi Dastcpsi ctcpsi Dxltcpsi ptcpsi DVxyi p0tcpsj DVxyj

Let the transfer function of the TCPS load flow controller be Tltcpsi s. It is
assumed that the transfer function of the TCPS stabilizer and feedback signal is
Tltcpsi s and the integral of deviation of the active power delivered along the
259
transmission line, DPij , where the TCSC is installed. When the TCPS stabilizer is
installed, from Fig. 6.10, it can have

1
D/i Tltcpsi s Tstcpsi s DPij Ttcpsi sDPij
s
D/i
Ttcpsi s
1 at c ps0 Ttcpsi s
T0tcpsi satcpsi DVxyi
6:88
atcpsj DVxyj
Let the state-space realization of Eq. (6.88) be

X_ TCPSi ATCPSi XTCPSi BTCPSi DVxyi BTCPSj DVxyj
6:89
D/i CTCPSi XTCPSi DTCPSi DVxyi DTCPSj DVxyj
6:90
By substituting Eq. (6.90) into Eq. (6.77), it can be obtained that

DITCSCij CTCSCij XTCSCi DTCSCij DVxyij
6:91
where
CTCSCij bTCPSi CTCPSi ;

bTCPSi DTCPSi
DTCSCij
0
0
DTCPSj
DTCPSi
Equations (6.89) and (6.91) are the linearized model of the TCPS with the
stabilizer installed between the ith and jth nodes in the N-machine power system.
6.1.2.5
General Linearized Model of the N-Machine Power System

Without the PSSs and Thyristor-Based FACTS Stabilizers
Installed
Similar to Eq. (5.52), network equations of M nodes in the N-machine power

system installed with thyristor-based FACTS stabilizers can be written as
2
6
6
6
4
I1
I2
..
.
IM
3 2
7 6
7 6
76
5 4
Y11
Y21
..
.
Y12
Y22
..
.
YM1
YM1
...
...
..
.
Y1M
Y21
..
.
. . . YMM
32
76
76
76
54
V1
V2
..
.
VM
3
7
7
7
5
6:92
260
where

Ii

g
ixi
v
; Vi xi ; Yij ii
bji
iyi
vyi

bij
; i; j 1; 2; . . .; M
gjj
6:93
Let the number of the SVCs, TCSCs, and TCPSs installed in the system be Lsvc ,
Ltcsc and Ltcps , respectively. Denote

T
DISVC DITSVC1 DITSVC2 . . . DITSVCLsvc ;
h
iT
DITCSC DITTCSC1j DITTCSC2j . . . DITTCSCLtcsc j ;
h
iT
DITCPS DITTCPS1j DITTCPS2j . . . DITTCPSLtcps j ;

T
DVSVC DVTSVC1 DVTSVC2 . . . DVTSVCLsvc ;
h
iT
DVTCSC DVTTCSC1j DVTTCSC2j . . . DVTTCSCLtcsc j ;
h
iT
DVTCPS DVTTCPS1j DVTTCPS2j . . . DVTTCPSLtcps j
6:94
When the SVCs, TCSCs, and TCPSs are not equipped with stabilizers,
Eqs. (6.42) and (6.48) are the linearized model of the ith SVC, Eqs. (6.63) and
(6.66) are that of the ith TCSC, and Eqs. (6.84) and (6.85) are that of the ith TCPS.
Hence, by using the denition of Eq. (6.94), the open-loop (without thyristor-based
FACTS stabilizers), state-space representation of Lsvc SVCs, Lt csc TCSCs, and Ltcps
TCPSs can be obtained, respectively, from Eqs. (6.42) and (6.48), Eqs. (6.63) and
(6.66), and Eqs. (6.84) and (6.85) as
X_ VSVC AVSVC XVSVC BVSVC DVSVC
DISVC CSVC XVSVC DSVC DVSVC bSSVC DaS
X_ LTCSC ALTCSC XLTCSC BTCSC DVTCSC bLTCSC DaSTCSC
DITCSC CTCSC XLTCSC DTCSC DVTCSC bTCSC DaSTCSC
X_ LTCPS ALTCPS XLTCPS BTCPS DVTCPS bLTCPS DaSTCPS
DITCPS CTCPS XLTCPS DTCPS DVTCPS bTCPS DaSTCPS
6:95
6:96
6:97
261
where
3
2
2
3
3
Dxlt c ps1
XVSVC1
Dxlt csc 1
7
6
6 X
6 Dx
7
7
6 Dxlt c ps1 7
lt csc 2 7
6 VSVC2 7
6
7
6
7; XLTCSC 6
7; XLTCPS 6
XVSVC 6
7;
..
..
..
6
6
7
7
7
6
4
4
5
5
.
.
.
5
4
xltcpsLtcps
Dxlt csc 1 Lt csc
XVSVCLsvc
3
2
2
2
3
3
Dastcps1
Das1
Dast csc 1
7
6
6 Da 7
6 Da
7
6 Dastcps2 7
s2 7
st csc 2 7
6
6
7
6
6
7
7; DaSTCPS 6
DaS 6
7
..
..
6 .. 7; DaSTCSC 6
7
7
6
4 . 5
4
5
.
.
5
4
DastcpsLtcps
DasLsvc
Dast csc Ltcsc
2
Without loss of generality and for the simplicity of expression, let the order of
arranging the network equations of Eq. (6.92) be (1) N nodes of generator terminals; (2) Lsvc nodes where Lsvc SVCs are installed; (3) Lt csc nodes where Lt csc
TCSCs are installed; (4) Ltcps nodes where Ltcps TCPSs are installed; and (5) other
nodes in the system. Linearized network equations of Eq. (6.92) can be written as
3 2
Ygg
DIxy
6 DISVC 7 6 Ysg
7 6
6
6 DITCSC 7 6 Ycg
7 6
6
4 DITCPS 5 4 Ypg
Yog
0
2
Ygs
Yss
Ycs
Yps
Yos
Ygc
Ysc
Ycc
Ypc
Yoc
Ygp
Ysp
Ycp
Ypp
Yop
32
3
Ygo
DVxy
7
6
Yso 7
76 DVSVC 7
6 DVTCSC 7
Yco 7
76
7
Ypo 54 DVTCPS 5
Yoo
DVO
6:98
where VO is the vector of the voltage at other nodes in the N-machine power
system.
The output equation of open-loop system of the SVC stabilizer is Eq. (6.40), that
of the TCSC stabilizer is Eq. (6.68), and that of the TCPS stabilizer is Eq. (6.87).
According to the notations in Eqs. (6.94)(6.98), the output equation of open-loop
system of Lsvc SVC stabilizers, Lt csc TCSC stabilizers, and Ltcps TCPS stabilizers
can be written, respectively, as
ySVC pSVC DVSVC pSVCO DVO
yTCSC dTCSC DaSTCSC cTCSC DXLTCSC pTCSC DVTCSC
yTCPS dTCPS DaSTCPS cTCPS DXLTCPS pTCPS DVTCPS
6:99
262
where
T
T
ySVC DP1j DP1j DPLsvc j ; yTCSC DP1j DP1j DPLt csc j ;

T
yTCPS DP1j DP1j DPLcps j ; dTCSC diagdt csc i ;
cTCSC diagct csc i ; dTCPS diagdtcpsi ; cTCSC diagctcpsi ;

2 0
psvc1
6
p0svc2
6
2
3
psvc1
6
6
0
6
6
7
psvc2
6
6
7
7; pSVCO 6
pSVC 6
.
6
6
7
.
6
4
5
.
0
0
6
6
psvcLsvc
6
4
pTCSC
6
6
6
6
6
6
6
6
6
6
6
6
4
2
pTCPS
6
6
6
6
6
6
6
6
6
6
6
6
6
4
3
7
7
7
7
7
7
p0svcLsvc 7
7;
7
7
0 7
7
5
..
.
0
..
ptcsc1
p0tcsc1
7
7
7
7
7
7
7
7
7
7
7
7
5
0
ptcsc2
p0tcsc2
..
.
ptcscLtcsc
p0tcscLtcsc
ptcps1
p0tcps1
0
ptcps2
p0tcsc2
..
.
ptcpsLtcps
p0tcpsLtcps
3
7
7
7
7
7
7
7
7
7
7
7
7
7
5
Substituting the second equation in Eqs. (6.95), (6.96), (6.97), and Eq. (6.29)
into Eq. (6.98), it can be obtained that

Cxy X Ba Du
0
22
Ygg Dgxy
66
Ysg
66
66
6
4
6
Ycg
6
Ypg
4
263
3
Ygo
7
Ysp
7
7
5
Ycp
Ypp DTCPS
Ygs
Ygc
Yss DSVC
Ysc
Ycs
Ycc DTCSC
Yps
Ypc
Y
Y
Y
Yop
og
os
oc

Y1o DV
Yoo DVO
Y0M
Y2o
33
Ygo
6 Y 77

6 so 7 7
7 7 DV
6
4 Yco 5 7
7 DV
O
7
Ypo 5
Yoo
2
6:100
where
2
Xg
3
2
DVxy
Cgxy
7
6 DV
6
SVC 7
6
6
DV6
7; Cxy 6
4 DVTCSC 5
4
7
6X
6 VSVC 7
X6
7;
4 XLTCSC 5
XLTCPS
DVTCPS
3
2
2
Dupss
0
7
6 Da
6
bSSVC
S
7
6
6
Du 6
7; B a 6
4 DaSTCSC 5
4
0
DaSTCPS
CSVC
CTCSC
CTCPS
3
0
7
7
7;
5
7
7
7
5
bTCSC
bTCPS

DVO Y1
oo Y2o DV
6:101
Cxy X Ba Du YM Y1o Y1

oo Y2o DV YM DV
By arranging the state equation (the rst equation) in Eqs. (6.95), (6.96), (6.97),
and (6.29) together according to the notation in Eq. (6.101), it can have
X_ Axy X Bp Du Bxy DV
6:102
where
2
6
6
Axy 6
4
2
6
6
Bp 6
4
Agxy
AVSVC
ALTCSC
ALTCPS
3
2
Bg
0
0
7
7
7;
5
7
6
7
6
7; Bxy 6
5
4
bLTCSC
bLTCPS
Bgxy
BVSVC
0
7
7
7
5
BTCSC
BTCPS
264
Deleting DV in Eq. (6.102) by substituting the second equation of Eq. (6.101)

into Eq. (6.102), it can be obtained that
X_ AX BDu
6:103
1
where A Axy Bxy Y1
M Cxy ; B Bp Bxy YM Ba . Equation (6.103) is the state
equation of open-loop system without the PSS and thyristor-based FACTS stabilizers installed.
Let the feedback signal of a PSS be the rotor speed of generator where it is
installed. Hence, the output equation of PSS can be written as
yPSS Dx 0
0 Xg cPSS Xg
6:104
By using the notation of Eq. (6.100), from Eq. (6.99), the rst equation in
Eq. (6.101) and (6.104), it can have
y C0 X D0 Du Pxy DV
6:105
where
2
yPSS
7
6y
6
6 SVC 7 0 6
y6
7; C 6
4 yTCSC 5
4
Pxy
yTCPS
2
0
6
pSVC
6
6
4
0
cPSS
7
6
7 0 6
7; D 6
5
4
cTCSC
3
0
cTCPS
0
0
0
7
7
7;
5
dTCSC
dTCPS
1
7
7 6 p
7 6 SVCO Yoo Y2o 7
7
76
5
5 4
0
pTCSC
pTCPS
Deleting DV in Eq. (6.105) by substituting the second equation of Eq. (6.101)

into Eq. (6.105), the output equation of the system about the PSS and
thyristor-based FACTS stabilizers can be obtained as
Dy CX DDu
6:106
1
0
where C C0 Pxy Y1
M Cxy ; D D Pxy YM Ba Du:
6.1.2.6
General Linearized Model of the N-Machine Power System

with the PSSs and Thyristor-Based FACTS Stabilizers
Installed
Equations (6.50) and (6.51) are the linearized model of an SVC with stabilizer
installed. Equations (6.70) and (6.72) are that of a TCSC with stabilizer installed.
265
Equations (6.89) and (6.91) are that of a TCPS with stabilizer installed. By using
the notation in Eqs. (6.94)(6.97), the linearized model of Lsvc SVCs, Lt csc TCSCs,
and Ltcps TCPSs with stabilizers installed can be obtained from Eqs. (6.50) and
(6.51), Eqs. (6.70) and (6.72), and Eqs. (6.89) and (6.91), respectively, as
X_ SVC ASVC XSVC BSVC DVSVC
DISVC CSVC XSVC DSVC DVSVC DSVCO DVO
6:107
X_ TCSC ATCSC XTCSC BTCSC DVTCSC
6:108
DITCSC CTCSC XTCSC DTCSC DVTCSC

X_ TCPS ATCPS XTCPS BTCPS DVTCPS
DITCPS CTCPS XTCPS DTCPS DVTCPS
6:109
where
2
6
6
XSVC 6
4
XSVC1
XSVC2
..
.
XSVCLsvc
7
6
7
6
7; XTCSC 6
5
4
XTCSC1
XTCSC2
..
.
7
6
7
6
7; XTCPS 6
5
4
XTCPS1
XTCPS2
..
.
3
7
7
7
5
XTCPSLtcps
XTCSCLtcsc
By substituting the second equation in Eqs. (6.107), (6.108), (6.109), and

Eq. (6.33) into Eq. (6.98), it can be obtained that
#
" 0

DV
Cp X
Yp Y1po
DVO
0
Y2o Yoo
22
3
Ygs
Ygc
Ygo
Ygg Dgp
66
7
Yss DSVC
Ysc
Ysp
6 6 Ysg
7
66
7
5
6
Ycs
Ycc DTCSC
Ycp
6 4 Ycg
6
Ypg
Yps
Ypc
Ypp DTCPS
4
Yog Yos Yoc Yop
33
Ygo
77
6 Y D

SVCO 7 7
6 so
DV
77
6
57
4
Yco
7 DV
O
7
Ypo
5
Yoo
2
6:110
where
3
2
Xg
Cgp
6 XSVC 7
CSVC
6
7
6
;C 4
X4
XTCSC 5 p
0
XTCPS
2
3
0
7
5
CTCSC
CTCPS
266

1
Y2o DV
DVO Yoo
0
1
Y2o DV Yp DV
Cp X Yp Y1po Yoo
6:111
By arranging the state equation (the rst equation) in Eqs. (6.107), (6.108),
(6.109), and (6.33) together, it can have
X_ Ap X B0p DV
where
2
6
6
Ap 6
4
2
6
6
B0p 6
4
6:112
3
Agp
ASVC
0
ATCSC
ATCPS
Bgp
BSVC
0
7
7
7;
5
3
7
7
7
5
BTCSC
BTCPS

X_ Aclose X
6:113
where Aclose Ap B0p Y1

p Cp . Equation (6.113) is the state equation of the
closed-loop system with the PSSs and thyristor-based FACTS stabilizers installed
in the N-machine power system.
6.2
Analysis and Damping Control of Thyristor-Based

FACTS Stabilizers Installed in a Multi-machine Power
System
6.2.1
Damping Torque Analysis in a Multi-machine Power

System
6.2.1.1
Damping Torque Contribution from a Stabilizer Derived

from the HeffronPhillips Model
Equation (5.43) is the HeffronPhillips model of the N-machine power system with
the PSSs installed. As this is a linear system satisfying the principle of superimposition, it only needs to consider the function of one PSS in the discussion. Now, if
6.2 Analysis and Damping Control
267
K1
0 I
s
(sM + D) 1
TPSS
K4
K5
K2
(K 3 + sTd0 )
(I + sTA ) K A
u pssk
Fk
K6
Fig. 6.11 HeffronPhillips model of a multi-machine power system installed with a PSS
only the kth PSS is considered, the HeffronPhillips model of the power system can
be obtained from Eq. (5.43) as
sDd xo IDx
0
0
sDE0q T01
6:114
0
1
0
sDE0fd T1
A DEfd TA KA K5 Dd K6 DEq Fk Dupssk
where Fi 0
... 0
. . . 0 T and Dupssk is the output stabilizing
"
the kth element
signal from the kth PSS. From Eq. (6.114), the HeffronPhillips model with the kth
PSS installed is shown in Fig. 6.11.
Similar to the procedure of damping torque analysis for a PSS installed in a
single-machine innite-bus power system as described in Chap. 2, from Fig. 6.11,
the transfer function matrix from the PSS stabilizing signal to the electromechanical
oscillation loops of all generators can be obtained as
DTPSS
M1 K2 I sTA K3 sTd0 KA K6 1 KA Fk FPSS s
Dupssk
6:115
Let the transfer function and feedback signal of the PSS be Tpssk s and Dyk ,
respectively, that is
Dupssk Tpssk sDyk
6:116
The feedback signal Dyk can always be reconstructed by using Dxj ; j

1; 2; . . .; N (see Sect. 6.2.1.2), that is
268
Dyk cj sDxj ; j 1; 2; . . .; N
6:117
If the low-frequency power oscillation of interests is associated with the ith

i n jxi , the electric torque contribution
oscillation mode of the power system, k
i
i n jxi can be
of the PSS at the complex angular oscillation frequency k
i
obtained from Eqs. (6.115) to (6.117) as
3
i T
i c k

pss1 k
pssk k
F
1 i Dx1
7
6

7
6
i )T
i Dy 6 Fpss2 ki Tpssk ki c2 ki Dx2 7
PSS k
pssk k
F
k
7
6
..
5
4
.
i T
i c k
i DxN
pssN k
pssk k
F
N
2
DTPSS
6:118
i ; j 1; 2; . . .N is the jth element of F

i . Hence, the damping
pssj k
PSS k
where F
torque contribution from the PSS to the electromechanical oscillation loop of each
generator is
i T
i c k

~pssj k
pssk k
DTdampj Ddampj Dxj ReF
j i Dxj
6:119
From the unied HeffronPhillips model of the N-machine power system

installed with a thyristor-based FACTS stabilizer of Fig. 6.5, the transfer function
matrix from the FACTS stabilizing signal to the electromechanical oscillation loops
of all generators can be obtained as
DTFACTS
FFACTS s)
Dufactss
M1 KP M1 K2 I sTA K3 sTd0 KA K6 1 I sTA Kq KA KV
6:120
Similarly, let the transfer function and feedback signal of the thyristor-based
FACTS stabilizer be Tfacts s and Dyk , respectively, that is
Dufactss Tfacts s)Dyk
6:121
The electric torque contribution of the thyristor-based FACTS stabilizer at the

i n jxi can be obtained from
complex angular oscillation frequency k
i
Eqs. (6.117), (6.120), and (6.121) as
3
i T
i
i Dx1
facts1 k
~ facts k
F
c1 k
i T
i
Dx2 7
6F
facts2 k
facts k
c2 k
i
7
i T
i Dy 6
FACTS k
facts k
F
7
6
.
k
..
5
4
i T
i
i DxN
factsN k
facts k
F
c k
2
DTFACTS
6:122
269
i ; j 1; 2; . . .; N is the jth element of F

i . Hence, the
FACTS k
factsj k
where F
damping torque contribution from the thyristor-based FACTS stabilizer to the
electromechanical oscillation loop of each generator is
i T
i c k

factsj k
facts k
DTdampj Ddampj Dxj ReF
j i Dxj
6.2.1.2
6:123
Damping Torque Contribution from a Stabilizer Derived

from the General Linearized Model of an N-Machine
Power System
The open-loop general linearized model of the N-machine power system is the state
equation of Eq. (6.103), that is
X_ AX BDu
6:124
where
3
2
3
Dupss1
Dupss
6 Dupss2 7
7
7
6X
6 Da
7
6
S
7
6 VSVC 7
6
7;
X6
7; Du 6
7; Dupss 6
.
7
6
4 XLTCSC 5
4 DaSTCSC 5
4 .. 5
XLTCPS
DaSTCPS
DupssN
3
2
3
3
2
2
Dastcps1
Das1
Dastcsc1
7
6
7
6 Da 7
6 Da
6 Dastcps2 7
s2 7
stcsc2 7
6
6
7
6
7
7
6
6
DaS 6 . 7; DaSTCSC 6
7
..
..
7; DaSTCPS 6
7
6
5
4 .. 5
4
.
.
5
4
DastcpsLtcps
DasLsvc
DastcscLtcsc
2
Xg
Equation (6.124) can be rearranged into the following form:

3 2
Dd_
0
6 Dx_ 7 4
5 A21
4
A31
DZ_
2
x0 I
A22
A32
3
2
32 Dd 3
0
0
X
7
6
7
6
A23 54 Dx 5
4 B2k 5Duk
k
A33
DZ
B3k
6:125
where Duk is the output stabilizing signal of the kth stabilizer installed in the power
system, which can be a PSS or a thyristor-based FACTS stabilizer, that is
Duk Dupssj ; j 1; 2; . . .; N; Duk Dasj ; j 1; 2; . . .; Lsvc ;
Duk Dastcscj ; j = 1,2,. . .; Ltcsc ; Duk Dastcpsj ; j 1; 2; . . .; Ltcps
270
A 21
A 22
1
s
u k
B3k
B 2k
0 I
s
A 32
A 31
A 23
1
s
A 33
Fig. 6.12 General linearized model of an N-machine power system
The general linearized model of the N-machine power system of Eq. (6.125) is
shown in Fig. 6.12. From the third row of Eq. (6.125), it can be obtained that
DZ sI A33
1
A31 Dd A32 Dx
!
B3k Duk
6:126
Substituting the above equation into Eq. (6.125), it can have

Dd_ x0 Dx
Dx_ A21 A23 sI A33 1 A31 Dd A22 A23 sI A33 1 A32 Dx
X
6:127
B2k A23 sI A33 1 B3k Duk
A21 sDd A22 sDx
Bk s)Duk
Hence, the general linearized model of Fig. 6.12 is shown by its compact form in
Fig. 6.13.
Let the transfer function and feedback signal of the kth stabilizer be Tstk s and
Dyk , respectively, that is
271
A 21 (s)
A 22 (s)
1
s
B k (s)
0 I
s
u k
Fig. 6.13 Compact form of general linearized model of an N-machine power system
Duk Tstk sDyk
6:128
From Fig. 6.13, Eqs. (6.117), and (6.128), the electric torque contribution from
the kth stabilizer can be obtained as
2
i T
i c k

k1 k
stk k
B
1 i Dx1
7
6

7
6
i T
i Dy 6 Bk2 ki Tstk ki c2 ki Dx2 7
k k
stk k
DTST B
k
7
6
..
5
4
.

BkN ki Tstk ki c ki DxN
N
i ; j 1; 2; . . .; N is the jth element of Bk k

i . Hence, the damping
where Bkj k
torque contribution from the kth stabilizer to the electromechanical oscillation loop
of each generator is (for j 1; 2; . . .; N)
i T
i c k

kj k
stk k
DTdampj Ddampj Dxj ReB
j i Dxj
6:129
For the kth stabilizer, the general linearized model of the multi-machine power
system of Eq. (6.125) can also be rearranged as (for j 1; 2; . . .; N)
3 2
Dd_ j
0
6 _ 7 4
D
x
k
4
j5
j
Aj1
DZ_ j
2
x0
dj
Aj2
32 Dd 3 2 0 3
j
0
7 6
7
6
Aj 54 Dxj 5 4 bjk 5Duk
Aj3
DZj
Bjk
6:130

DZj sI Aj3 1
hx
i
Aj1 Aj2 Bjk Tstk scj s Dxj
6:131
The output equation of the system about the feedback signal of the stabilizer can
be written as
272
Ddj

6
7
Dyk c1kj c2kj C3kj 4 Dxj 5 djk Duk
DZj
6:132
By using Eqs. (6.117), (6.128), and (6.131), Eq. (6.131) becomes

Dyk
x
1 x0
c1kj c2kj Dxj C3kj sI Aj3
Aj1 Aj2 Dxj
s
s
C3kj sI Aj3 1 Bjk djk Tstk scj sDxj
0
6:133

cj s

x

Dyk x0
0
c1kj c2kj C3kj sI Aj3 1
Aj1 Aj2
Dxj
s
s
6:134
C3kj sI Aj3 1 Bjk djk Tstk scj s

Hence, nally, it can have
x0 c1kj c2kj C3kj sI Aj3 1 xs0 Aj1 Aj2
cj s s
C3kj sI Aj3 1 Bjk djk Tstk s
6:135
6.2.2
Selection of Installing Location and Feedback Signal

of a Stabilizer in a Multi-machine Power System
6.2.2.1
Selection of Installing Location of a PSS by Participation

Factor and Sensitivity Index
Selection of installing location and feedback signal of a stabilizer is considered

before a stabilizer is installed in a multi-machine power system. At this stage, both
the structure and parameters of the stabilizer are not decided. The selection is based
on the open-loop linearized model of the power system to predict the effectiveness
of the stabilizer. The participation factor, the sensitivity index, the residue, and the
damping torque analysis are among the most popular techniques for the selection.
Consider the general linearized open-loop system of Eqs. (6.130) and (6.132).
i
Denote the right and left eigenvectors of the state matrix with respect to the mode k
to be (see Sect. 2.2.1 and 5.2.1.1)
2
3
v1i

Ti w
vi 4 v2i 5; w
1i
v3i
2i
w
T3i
w
6:136
273
The following participation factor (see Eq. (5.60)

2i v2i j
jw
pji T
i vi
w
6:137
measures how much the ith oscillation mode and the variation of rotor speed of the
jth generator are connected. If the connection is the highest among all generators as
measured by the above-dened participation factor, the jth generator is chosen to be
the best installing location of a PSS to suppress the power system oscillation
i . The general procedure of the selection is to
associated with the oscillation mode k
arrange the state matrix of the system according to Eq. (6.125), that is
2
0
A 4 A21
A31
x0 I
A22
A32
3
0
A23 5
A33
6:138
Denote the right and left eigenvectors of the above state matrix with respect to
i to be
the mode k
2
3
v1i
T
Ti w
1i
vi 4 v2i 5; w
v3i
T2i
w
T3i
w
6:139
The participation factors of all generators can be calculated as

vj2i uj2i

pji T for j 1; 2; . . .; N
v2i
w
6:140
2i
The installing location of the PSS is the generator with the highest participation
factor.
The sensitivity index for the selection of installing location of the PSS is
computed by adding an increment of articial damping in the electromechanical
oscillation loop of each generator in the power system, Ddj Dxj (for j 1; 2; . . .; N).
i , is
i , for the oscillation mode of interests, k
If the maximum improvement, Dk
achieved by the increment on the jth generator, it is selected as the best installing
location of the PSS. The physical explanation of the sensitivity index is based on
the understanding that the PSS is equivalent to an addition of damping torque in the
electromechanical oscillation loop of generators. Hence, if the addition of damping
torque on the jth generator achieves the maximum improvement of the oscillation
mode of interests, it should be the most effective installing location of the PSS as far
as the improvement of the oscillation mode of interests is concerned. The sensitivity
index in fact is the sensitivity of the oscillation mode of interests to the coefcient
of the added extra damping torque in the electromechanical oscillation mode, Ddj .
That is
274

ij Dki @ ki
S
Ddj @dj
6:141

@A
ij @ ki w
Ti
S
vi w
1i
@dj
@dj
2i
w
2
0

T3i 4 0
w
0
0
1
0
32 3
v1i
0
2i v2i 6:142
0 54 v2i 5 w
v3i
0
Equation (6.142) shows that in fact, the participation factor is the normalized
sensitivity index. Both methods are equivalent. Application of the sensitivity index
often only considers the real part of the oscillation mode of interests, as the
objective of installing stabilizer is to improve the damping of the oscillation mode.
i n jxi . The sensitivity index used is
Let the oscillation mode be k
i
Sijn
Dni
Ddj
6:143
Though Eqs. (6.119) and (6.129) show that a PSS supplies damping torque to the
electromechanical oscillation loop of all generators in the power system, usually the
amount of damping torque contribution to the generator where it is installed is much
greater than that of other generators. Hence, participation factor and sensitivity
index, though they only tell how much a generator is connected to the oscillation
mode of interests, can be used to select the installing location of the PSS. In the case
of a thyristor-based FACTS stabilizer, however, they cannot be applied directly.
This is because how the thyristor-based FACTS stabilizer distributes its damping
torque contribution among all generators also plays a very important role in
determining how much it affects the oscillation modes of interest.
6.2.2.2
Selection of Installing Locations and Feedback Signals

of a Stabilizer by Damping Torque Analysis [2]
Equations (6.119) and (6.129) demonstrate that a stabilizer (the PSS or

thyristor-based FACTS stabilizer) installed in a power system contributes a
damping torque, Ddampj Dxj , to the electromechanical oscillation loop of every
generator in the system. By using the sensitivity index dened in Eqs. (6.141) or
(6.143), the addition of damping torque, Ddampj Dxj (for j 1; 2; . . .; N), on each
generator affects the oscillation mode of interests. Hence, the total improvement of
i by the stabilizer can be obtained as
the oscillation mode of interests k
i
Dk
N
X
j1
ij ; or Dni
Ddampj S
N
X
j1
Ddampj Sijn
6:144
275
Generator 1
Si1
Ddamp1
Generator 2
Ddamp2
Si2
The
oscillation
mode of
interests
A
stabilizer
SiN
DdampN
Generator N
Damping torque contribution

to each generator
Damping torque contribution is

converted to the improvement
of the oscillation mode of
interests
Fig. 6.14 The oscillation mode of interests as affected by a stabilizer
where Ddampj is given by Eq. (6.119) for a PSS and Eq. (6.123) for a thyristor-based
FACTS stabilizer based on the HeffronPhillips model as well as by Eq. (6.129) for
the PSS or thyristor-based FACTS stabilizer based on the general linearized model
of the N-machine power system.
Equation (6.144) is shown in Fig. 6.14. It presents a clear and full picture on how
the stabilizer (the PSS or thyristor-based FACTS stabilizer) contributes the damping
torque to every generator and then how the contribution of the damping torque is
converted to the influence on the oscillation mode of interests. Damping torque
analysis in the multi-machine power system of Eqs. (6.119), (6.123), and (6.129)
gives the left-hand half picture of Fig. 6.14 on how the stabilizer distributes the
damping torque to all generators. Participation factor or sensitivity index gives
equivalently the right-hand half picture of Fig. 6.14 on how the oscillation mode of
interests is connected to generators.
Based on Fig. 6.14 or Eq. (6.144), the following indices of the damping torque
analysis (DTA) can be proposed for the selection of installing location and feedback
signal of the stabilizer:
DTA
N
X

i c k

F
pssj k
j i Sij for a PSS using HeffronPhillips model 6:145
j1
DTA
N
X

i c k

F
fsctsj k
j i Sij
j1
6:146
276
(for a thyristor-based FACTS stabilizer using HeffronPhillips model)

DTA
N
X

i c k

B

kj k
j i Sij
6:147
j1
(for a PSS or thyristor-based FACTS stabilizer using general linearized model)

where cj s xs0 c1kj c2kj C3kj sI Aj3 1 xs0 Aj1 Aj2 which is given by
Eq. (6.135).
ij is replaced by Sijn so as to focus the
In the above DTA indices, often S
selection on the influence of the stabilizer on the damping of the oscillation mode of
interests.
6.2.2.3
Residue Index and Its Connection with the Damping Torque

Analysis [3]
Consider the kth stabilizer to be installed in a multi-machine. The general linearized

model of the system is (see Eq. (6.125))
2
Dd_
0
7
6
6 Dx_ 7 6
5 4 A21
4
DZ_
A31

Dyk CT1k
CT2k
32 Dd 3 2
3
x0 I 0
0
7 6
76
7
7
A22 A23 56
4 Dx 5 4 B2k 5Duk
A32
A33
DZ
2
3
Dd

7
T 6
Dx
C3k 4
5
DZ
B3k
6:148
where Dyk is the feedback signal of the stabilizer. If the right and left eigenvectors
i , are vi and w
Ti ,
of the state matrix with respect to the oscillation mode of interests, k
respectively, the controllability, observability, and residue index of the stabilizer are
(see Sect. 2.2.1)
2
3
0

Ti 4 B2k 5; cik CT1k
bik w
B3k
CT2k

ik bikcik
CT3k vi ; R
6:149
i is
Figure 6.15 shows that cik measures how much the oscillation mode k
i .
observed in Dyk and bik measures how much the stabilizing signal Duk affects k

Hence, the residue index Rik measures how much the stabilizer is connected with
i . It can be used as an index for the selection of
the oscillation mode of interests k
the installing location and feedback signal of the stabilizer.

Fig. 6.15 Conguration of
the system in modal
decomposition
277
1
s i
bik
cik
Channels associated with other

eigenvalues of the system
u k
+
y k
stabilizer
In fact, the damping torque analysis and the residue introduced above for the
selection of installing location and feedback signal of the stabilizer are equivalent.
This equivalence can be proved as follows.
3
z1 0ek1 t
Dd
6 z2 0ek2 t 7
7
4 Dx 5 V6
7
6
..
5
4
.
DZ
zn 0ekn t
2
6:150
Hence,
Dxj s
N
X
v2ij zi 0
i
sk
6:151
i1
where
v2ij is the jth element of vi2 in Eq. (6.139). While

Dyk CT1k
CT2k
2
3
Dd

CT3k 4 Dx 5 CT1k
DZ
3
z1 0ek1 t
k2 t 7
6
6 z2 0e 7
CT3k V6
7
..
5
4
.
2
CT2k
zn 0ekn t
Hence,
Dyk s
N
X
cik zi 0
i
sk
i1
6:152
278

PN c ik zi 0
i PN c ik zi0
s k
i
Dyk s
i1 sk
i1 ski
Dxj s PN v 2ij zi 0 s k

i PN v 2ij zi 0
i1 ski
i1 ski
6:153
i and using Eq. (6.117), it can be obtained

In the above equation, by letting s k
that
i
cj k
i cik zi 0
cik
Dyk k
i v2ij zi 0 v2ij
Dxj k
6:154
On the other hand, from Eqs. (6.139) and (6.148), it can have
T1i
w
T2i
w
2
0

T3i 4 A21
w
A31
x0 I
A22
A32
3
0
T
i w
1i
A23 5 k
A33
T2i
w
T3i
w
6:155

i
T3i A33 w
T3i k
T2i A23 w
w
6:156
Solution of Eq. (6.156) is

i I A33 1
T2i A23 k
T3i w
w
6:157
Hence, from Eqs. (6.148), (6.149), and (6.157), it can have

2
3
0
6
7 T

Ti 4 B2k 5 w
1i
bik w
3
0
6
7
T2i B2k w
T3i B3k
T2i w
T3i 4 B2k 5 w
w
B3k
B3k
X
1
T
i I A33 B3k w
i
i
2i B2k A23 k
T2i Bk k
2ij Bkj k
w
w
6:158

2
Sij
0
@ 4 A21
A31
T
i
@k
i
w
@dj
3
0
x0 I
A22 diagdj A23 5
A32
A33
vi w
2ij v2ij
@dj
6:159
Hence, from Eqs. (6.154), (6.158), and (6.159), the residue index can be
obtained as
ik
R
bikcik
279
i c k

kj k
2ij B
w
j i v2ij
i c k

ij B
kj k
S
j i
6:160
This proves that the DTA index of Eq. (6.147) for the selection of installing
ik j.
location and feedback signal is equivalent to the residue index jR
6.2.3
Selection of Robust Installing Locations and Feedback

Signals of a Stabilizer by an Eigensolution-Free
Method
6.2.3.1
Robust Selection of Installing Locations and Feedback Signals

of a Stabilizer [4]
A good design of a stabilizer must ensure the maximum effectiveness of the stabilizer. In addition, the robustness of the stabilizer to the variations of power system
operation conditions is also an important factor to be considered in the design. In
Sect. 4.3.2.3, the issue of robust design of the stabilizer is introduced in the
selection of control functions of a UPFC to add the stabilizing control. In this
section, it is discussed how the robustness of the stabilizer is considered at the stage
of selecting the installing location and feedback signal.
Let the set of the candidate installing locations and feedback signals of the
stabilizer in a multi-machine power system be Uu and that of power system
operating conditions Xl. The effectiveness of the stabilizer is the function of
and , C(, ). If only the maximum effectiveness of the stabilizer to damp a power
oscillation is considered, the criterion of selecting the installing location and
feedback signal of the stabilizer is
MaxCu; l0 ;
u
l0 2 Xl
6:161
In the discussion in previous sections, only the above criterion is considered.

In order to consider the robustness of the stabilizer to the variations of power
system operating conditions, the following two criteria should be added.
Maxmin Cu; l
6:162
Minmax Cu; l min Cu; l
6:163
1. The criterion of Eq. (6.162) means that the selection of installing location and
feedback signal is made at the operating condition where the stabilizer is least
effective. This selection guarantees that when the stabilizer operates at other
280
operating conditions, it is more effective. Hence, the selection is robust to the

variations of system operating conditions.
2. If the damping contribution from the stabilizer changes signicantly, on the one
hand, with the variations of power system operating conditions, the control
could be overstrong at some operating conditions, which would pose much
unwanted influence on other modes in the power system. Hence, the criterion of
Eq. (6.163) requires that the change of damping contribution from the stabilizer
is minimum with the variations of power system operating conditions. However,
this criterion should be applied jointly with that of Eq. (6.162), since failing to
meet the requirement of the effectiveness obviously is not a proper selection.
Obviously, in order to apply the criteria of Eqs. (6.162) and (6.163), the effectiveness of the stabilizer has to be examined over the range of all system operating
conditions. If a method dependent on the eigensolution of system state matrix, such
as the residue index, is used, it could mean a heavy computational burden. In the
following section, an equivalent residue index is derived which can be used for the
eigensolution-free selection of robust installing location and feedback signal of the
stabilizer.
6.2.3.2
An Equivalent Residue Index and Eigensolution-Free Selection

of Robust Installing Location and Feedback Signal
of the Stabilizer
Consider the kth stabilizer to be installed in the N-machine power system. The
linearized model of the system is given by Eqs. (6.130) and (6.132) as follows:
32 Dd 3 2
3
j
0
7
7 6
6
76
7
7
6 Dx_ j 7 6
dj
Aj 56
5 4 kj
4
4 Dxj 5 4 bjk 5Duk
Aj1 Aj2 Aj3
Bjk
DZj
DZ_ j
3
2
Ddj
7

6
7
Dyk c1kj c2kj C3kj 6
4 Dxj 5
DZj
2
Dd_ j
x0
6:164
Let the right and left eigenvectors of the state matrix, respectively, be
2
3
v1i

Ti w
vi 4 v2i 5; w
1i
v3i
2i
w
T3i
w
6:165
281
According to the denition of the right and left eigenvectors, it can have
1i
w
2i
w
0
6
4 kj
Aj1
x0
dj
Aj2
3
2
0
0
x0
6
7
dj
Aj 5 k
T3i 4 kj
1i
w
i w
Aj1 Aj2 Aj3
32 3
2 3
v1i
0
v1i
76 7 6 7
Aj 54 v2i 5 ki 4 v2i 5
Aj3
v3i
v3i
2i
w
T3i
w

6:166

i I Aj3 1 w
Ti3 Aj k
i2
w
x0
vi1 vi2
ki

i I Aj3 1 Aj1 x0 Aj2 vi2
vi3 k
i
k
6:167
From Eq. (6.167), controllability and observability index can be obtained,

respectively, as
2

bik w
1i
2i
w
6
7
T3i 4 bjk 5
w
Bjk
i I Aj3 1 Bjk w
i w
bi k
2i K
i2
bjk Aj k
2 3
v1i

6 7
cik c1kj c2kj C3kj 4 v2i 5
v3i

x0
x0
1

i v2i
ci k
c1kj c2kj C3kj ki I Aj3 Aj1 Aj2 v2i K
ki
ki
6:168
Hence, the equivalent residue index is
i K
i vi2 w
ik bikcik K
bi k
ci k
i2
R
6:169
In the selection of installing location and feedback signal of the stabilizer to be

installed in the multi-machine power system, the residue is calculated for comparison among various candidate locations and feedback signals. For example, if
there are two candidate installing locations or feedback signals, A and B, and
ikA j [ jR
ikB j, then A is considered to be better than B to be the installing location
jR
282

R
ikA
or the feedback signal of the stabilizer. Therefore, it is the ratio, , not the
RikB
values of the residue, that determines the selection.
On the other hand, in system state equation of Eq. (6.164), if the installing
location or the feedback signal of the stabilizer is different, the control and the
output vectors are not the same, but the open-loop state matrix is unchanged,
provided that the operating point of the system remains the same. That is,
vi2A vi2B ; w
i2A w
i2B
6:170
Therefore, from Eqs. (6.169) and (6.170), it can be obtained that

i K
i vi2A w
biA k
ciA k
ikA j K
i2A
jR

ikB j K
i K
i vi2B w
biB k
ciB k
jR
i2B

i K
i jvi2A w
i K
i
K
biA k
biA k
ciA k
ciA k
i2A j K

i K
i jvi2B w
i K
i
biB k
biB k
ciB k
ciB k
i2B j K
K
6:171
bi ki K
ci ki j can replace jR
ik j to be the index to
Equation (6.171) shows that jK
measure the effectiveness of the stabilizer so as to select the installing location and
feedback signal. Since in most cases, the
mode of
oscillation

interest is lightly
i jxi , so K
K
K
bi jx K
bi k
ci k
ci jx , the index can be
damped, i.e. k
i
i
i
i
calculated without knowing the eigensolution of the open-loop state matrix.
Therefore, the selection can be made eigensolution-free with the following criteria:

bi jxi K
ci jxi j; for all l 2 Xl ; for all u 2 Uu
Max MinjK
u
l

bi jxi K
bi jxi K
ci jxi j MinjK
ci jxi j;
Min MaxjK
u
6:172
for all l 2 Xlg; for all u 2 Uu
6.2.4
Stabilizer Design in a Multi-machine Power System

Considering Robustness and Interaction of Stabilizers
6.2.4.1
Selection of an Operating Condition for the Robust

Coordinated Design of Multiple Stabilizers in a Multi-machine
Power System
Section 4.3.2.3 discusses the selection of an operating condition for the design of a
robust stabilizer in a single-machine innite-bus power system by using the phase
compensation method. The selection strategy presented in Sect. 4.3.2.3 is to choose
283
an operating condition where it is predicted that the stabilizer to be installed will be

least effective. Hence, once the stabilizer is designed at the selected operating
condition to be effective in damping power system oscillation, its effectiveness is
ensured when the system operates at other operating conditions.
In a multi-machine power system installed with multiple stabilizers, such a
strategy to select an operating condition for the design of robust stabilizers can be
similarly applied by using the damping torque analysis as follows.
Consider the kth stabilizer installed in the system to improve the damping of the
i n jxi , and use the following sensitivity index given by
ith oscillation mode, k
i
Eq. (6.143)
Dni
N
X
Ddampj Sijn
6:173
j1

Dni
N
X
Ddampj Sijn
j1
N
X
i c k

kj k
ReB
j i Sijn
6:174
j1
which gives the prediction of the amount of damping provided by the kth stabilizer
to the ith oscillation mode, that is, the effectiveness of the stabilizer. Therefore, if it
is assumed that the set of all known candidate operating conditions of the
multi-machine power system is X0 , the following index can be used to predict the
effectiveness of the kth stabilizer to damp the ith oscillation mode at an operating
condition:
Wik m
N
X
i c k

kj k
ReB
j i Sijn ; lik m 2 X0
6:175
j1
If Wik mt minWk ; lik m 2 X0 , it is determined that the kth stabilizer is

m
least effective in damping the ith oscillation mode at the operating condition
lik mt 2 X0 . Hence, lik mt 2 X0 can be chosen as the operating condition to
design the kth stabilizer. An effective design of the stabilizer at lik mt 2 X0 can
ensure the effectiveness of the stabilizer at all operating conditions lik mt 2 X0 .
Obviously, this selection above may result in different operating conditions for
different stabilizers and oscillation modes. So, it may be difcult or impossible to
select a common operating condition at which all stabilizers can be designed
simultaneously in coordination. In other words, if they are designed in coordination
at a common operating condition simultaneously, their robustness may not be
ensured due to the difference in the changing pattern of their effectiveness to the
variations of system operating conditions. This problem can be solved by designing
284
the multiple stabilizers at two stages. The rst stage of the design is to set the phase
of stabilizers one by one in a sequence according to the selection of robust operating conditions of stabilizers. At the second stage, gains of all stabilizers are tuned
simultaneously in coordination to avoid the problem of eigenvalue drift.
For simplicity of expression, the following conventional form of the transfer
function of stabilizers is used:
Tstk s Kk
sTwk
1 1 sTk2 1 sTk4
Kk Tk s
1 sTwk 1 sTk 1 sTk1 1 sTk3
6:176
It is assumed that the kth stabilizer is designed mainly to improve the damping of
i n jxi . Without loss of genthe ith oscillation mode in the power system, k
i
erality, it is assumed that at the operating condition of the kth stabilizer selected by
the use of Eq. (6.175), Si1n
Si2n

SiNn . A compensation angle /k is
chosen to satisfy
90 [ /k ukj [ 0 ;
j 1; 2; . . .; N
6:177
i c k

kj k
where ukj is the phase of B
j i . If such a compensation angle cannot be
obtained, the selection of the compensation angle /k can be made among the rst
(N 1) machines to satisfy
90 [ /k ukj [ 0 ;
j 1; 2; . . .; N 1
6:178
This procedure can be continued until such a compensation angle is obtained.

i jTk j\ / .
k k
Then, Tk s is set to have a phase angle /k , that is, T
k
According to Fig. 6.14, as far as the setting of the transfer function of the stabilizer is
i can be expressed as
concerned, the damping contribution by the kth stabilizer to k
Dni k
N
X
j1
N
X
Ddampj Sijn
N
X
i c k

kj k

ReB
j i Sijn ReDTstk ki
j1
Sijn Hkj Tk cos bkj DKk ;
6:179
j 1; 2; . . .; N
j1
i
i T
i Hkj Tk \b ; j 1; 2; . . .; N and b / u .
stk k
kj k
cj k
where B
kj
kj
k
kj
The procedure above in selecting the compensation angle /k ensures that
i so that
0 [ bkj [ 90 is tenable to the machines which are more sensitive to k
they are provided with a positive synchronizing torque. Equation (6.179) shows the
damping contribution to the ith oscillation mode by the kth stabilizer set at the
285
selected operating condition. If there are total L stabilizers considered, the total
damping contributed from L stabilizers to the ith oscillation mode can be obtained
from Eq. (6.179) as
Dni
L
X
k1
Dni k
L X
N
X
Sijn Hkj Tk cos bkj DKk ;
j 1; 2; . . .; N 6:180
k1 j1
Let the target damping of the oscillation modes by the coordinated design of L
stabilizers be ni ; i 1; 2; . . .; L. An objective function can be formed as
JK
L
X
Qi ni K ni 2
6:181
i1
where K K1 K2 . . . KL T and Qi is the weight for the assignment of the ith

oscillation mode. The simultaneous tuning of gains of L stabilizers can be carried
out by using a method of optimization to nd the solution of the objective function
dened by Eq. (6.181). This simultaneous tuning is conducted at the selected
operating condition of each stabilizer. At the selected operating condition of the rst
stabilizer, the initial values of all gains are simply taken to be zero. From the
selected operating condition of the second stabilizer on, the gain values set by the
previous tuning are taken to be the initial values to start the current tuning. Also in
the tuning process, the following constraint is checked:

6:182
jni Kj
ni ; i 1; 2; . . .; L
If Eq. (6.182) is satised for any kl ; l 1; 2; . . .; L, which is supposed to be
damped by the kth stabilizer, Ql 0 is set in Eq. (6.181). This arrangement frees
the eigenvalue drift towards the good direction, the left half complex plane over
the target value. Therefore, a unique nal solution of gain tuning will be much
easier to be found.
6.2.4.2
Non-negatively Interactive Design of a Stabilizer

in a Multi-machine Power System [5]
For the design of multiple stabilizers (the PSSs or FACTS-based stabilizers) in a

multi-machine power system, the techniques of sequential setting require the stabilizers to be designed one by one in a sequence. However, the stabilizer set later
may move the oscillation modes damped by the stabilizers designed previously.
286
This problem of eigenvalue drift caused by the interactions among the multiple
stabilizers has motivated the development of the coordinated tuning of multiple
stabilizers. By coordinated tuning, the parameters of all interactive stabilizers are set
simultaneously to achieve an accurate assignment of all oscillation modes of
interest in the power system so that the interactions among the stabilizers are fully
taken account of.
However, in the design of multiple stabilizers, only an approximate assignment
of oscillation modes is needed as long as the damping of the oscillation modes is
improved to meet a certain target value of damping. If the interactions among the
stabilizers are positive, in fact, extra benets of damping provision can be obtained.
Hence, the positive interactions should be encouraged instead of being strictly
constrained by the coordinated tuning. While what ought to be avoided is the
negative interactions among multiple interactive stabilizers, which may result in the
provision of negative damping to some oscillation modes, at the same time that
positive damping is contributed by the design of the stabilizer. Furthermore, in
practice, it is not always the case that when a new stabilizer is put into a power
system, all old interactive stabilizers are adjustable so that the coordinated tuning
can be implemented. As long as the design of this new stabilizer can ensure no
provision of negative damping to all interactive oscillation modes in the power
system, it can be set without the coordinated tuning with other old stabilizers. In this
section, it is discussed how such a non-negatively interactive stabilizer can be
designed.
From Eq. (6.179), the damping provided by a stabilizer (the kth stabilizer) in a
multi-machine power system can be denoted as
Dni k
N
X
i c k

kj k

ReB
j i Sijn ReDTstk ki
j1
Re
N
X
)
i DT
i
kij k
stk k
Sijn H
6:183
i
ki DT
stk k
ReC
j1
P
i jC
i . If the target oscillation mode for
ki N Sijn Hkij k
ki j\aki k
where C
j1
0 n jx0
the
installation
the
kth
stabilizer
is
k
and
0

of
0 K k T
0 \b k
0 , from Eq. (6.183) the damping improvement of the
k k
stk k
T
k0
target oscillation mode by the stabilizer can be obtained as

0 jC
0 cosb k

k0 DT
k0 jDKk T
k k
stk k
Dn0 ReC
k0 0 ak0 k0
6:184
287
i n jxi ; i 6 0, in the power system, the

For other oscillation modes, k
i
damping increment of the ith oscillation mode due to the installation of the kth
stabilizer is

i jC
i cosb k

ki DT
ki jDKk T
k k
stk k
Dni ReC
ki i aki ki
6:185
To ensure a positive damping provision by the stabilizer to the target oscillation

mode and at the same time, the imposition of a non-negatively interaction upon the
ith oscillation mode, it should have
0 ak0 k
0 [ 0; cosb k

cosbk0 k
ki i aki ki
0
6:186
Therefore, a successful design of non-negatively interactive stabilizer is equivi , onto the rst quadrant in the complex plane,
ki k
alent to shifting all the vectors, C
i n jxi ; i 0; 1; . . . are oscillation modes associated with the interwhere k
i
active stabilizers in the power system. From Eq. (6.184) to (6.186), obviously, it
can be seen that the following two factors decide whether the shifting may be
successful:
0 and aki k
i ;
1. the difference between the angles ak0 k
2. the setting of the phase angle of the stabilizer.
i is onto the positive real axis so that
ki k
The most successful shifting for C

cosbki ki aki ki 1; i 0; 1; . . ..
i [ ak0 k
0 , to achieve the maximum efciency of the
In the case that aki k
stabilizer to damp its target oscillation mode, according to Eq. (6.184), the phase of
0 ak0 k
0 so that C
0 is shifted onto
k0 k
the stabilizer could be set to be bk0 k

0 may
k0 k
the real axis with cosbk0 k0 ak0 k0 1. As shown in Fig. 6.16a, C
also be moved into the rst quadrant. This results in a design of an effective and
non-negatively interactive stabilizer.
i \ak0 k
0 as shown in Fig. 6.16b, the priority of shifting may
However, if aki k
0 so as to move it to the rst quadrant and in the same
k0 k
have to be given to C

time to shift Ck0 k0 into the rst quadrant instead of onto the real axis. Thus, the
design may not be most efcient as far as the damping of the target oscillation
0 , is concerned. However, the stabilizer is positively interactive with k
i .
mode, k

In both cases above, if the angle between Ck0 k0 and Cki ki is large and even
0 aki k
i [ 90 as shown in Fig. 6.16c, there is the danger that when
ak0 k
0 is shifted into the rst quadrant, C
i may not be able to be moved into
k0 k
ki k
C
288
Fig. 6.16 Non-negatively
interactive design of a
stabilizer.
i [ ak0 k
0 ,
a aki k
i \ak0 k
0 ,
b aki k
0 aki k
i [ 90
c ak0 k
(a)
Cki ki ( i ) + ki ( i )
Ck 0 k 0 ( 0 ) + k 0 ( 0 ) = 0
k 0 ( 0 )
Cki ( i )
C k 0 ( 0 )
(b)
Ck 0 k 0 ( 0 ) + k 0 ( 0 )
Cki ki ( i ) + ki ( i ) = 0
x
ki ( i )
C k 0 ( 0 )
Cki ( i )
(c)
Cki ki ( i ) + ki ( i )
y
C ki ( i )
Ck 0 k 0 ( 0 ) + k 0 ( 0 ) = 0
x
k 0 ( 0 )
C k 0 ( 0 )
289
the rst quadrant in the same time which could result in a negative damping
provision to the ith oscillation mode. This means a negative interaction between the
stabilizer being designed and the stabilizer which damps the ith oscillation mode.
6.3
An Example Two-Area Four-Machine Power System
6.3.1
Linearized Model
6.3.1.1
System Parameters and Operating Conditions
Conguration of a four-machine eleven-node example power system is shown in

Fig. 6.17.
Line parameters are given in Table 6.1. Parameters of generators and the AVRs
are given in Table 6.2.
G1 1
L7
10
11
G3
L9
C7
C9
G2
G4
Fig. 6.17 Conguration of a four-machine example power system
Table 6.1 Network

parameters
Node
Node
Resistance
Reactance
Susceptance
1
2
3
4
5
6
7
7
8
8
9
10
5
6
11
10
6
7
8
8
9
9
10
11
0.0000
0.0000
0.0000
0.0000
0.0025
0.0010
0.0110
0.0110
0.0110
0.0110
0.0010
0.0025
0.0167
0.0167
0.0167
0.0167
0.0250
0.0100
0.1100
0.1100
0.1100
0.1100
0.0100
0.0250
0.0000
0.0000
0.0000
0.0000
0.04375
0.0175
0.1925
0.1925
0.1925
0.1925
0.0175
0.04375
290
Table 6.2 Generator parameters

TJ
Xd
Xq
X0d
T0d0
D
M
KA
TA
G1
G2
G3
G4
13.0
1.8000
1.7000
0.3000
8.000
13.0
1.8000
1.7000
0.3000
8.000
12.35
1.8000
1.7000
0.3000
8.000
12.35
1.8000
1.7000
0.3000
8.000
0
6.5
50
0.1
0
6.5
50
0.1
0
6.175
50
0.1
0
6.175
50
0.1
Table 6.3 Results of load flow computationnode voltage and node injection power
Node
Magnitude of
voltage
Phase of voltage (deg.)
Active
power
Reactive
power
G1
G2
G3
G4
5
6
7
8
9
10
11
1.0300
1.0100
1.0300
1.0100
0.9923
0.9434
0.8989
0.8636
0.8799
0.9320
0.9869
0.5362
0.3616
0
0.1910
0.4218
0.2388
0.0792
0.2000
0.4834
0.3153
0.1205
7.0000
7.0000
7.3319
7.0000
0.0000
0.0000
9.6700
0.0000
17.6700
0.0000
0.0000
2.7299
4.4625
3.1059
5.1643
0.0000
0.0000
1.0000
0.0000
1.0000
0.0000
0.0000
Initial conditions of the example power system for the load flow computation are
VG1 1:03; VG2 1:01; VG3 1:03\0; VG4 1:03;
PG1 7:0; PG2 7:0; PG4 7:0;
Load L7 9:67 j1:00; Load L9 13:67 j1:00:
Results of load flow computation are given in Tables 6.3 and 6.4. Based on the
results of load flow, initial values of state variables can be calculated as follows.
6.3 An Example Two-Area Four-Machine Power System
291
Table 6.4 Results of load flow computationline power and power loss
Starting node
(SE)
Ending node
(EN)
Power at SN
Power at EN
Power loss
1
2
3
4
5
6
7
7
8
8
9
10
5
6
11
10
6
7
8
8
9
9
10
11
7.00 + j2.73
7.00 + j4.46
7.33 + j3.10
7.00 + j5.16
7.00 + j1.82
13.87 + j3.88
1.98 + j0.21
1.98 + j0.21
1.93 j0.19
1.93 j0.19
13.93 j2.04
7.18 j0.67
7.00 j1.84
7.00 j3.33
7.33 j2.11
7.00 j3.93
6.86 j0.57
13.63 j1.58
1.93 + j0.04
1.93 + j0.04
1.87 + j0.44
1.87 + j0.44
14.18 + j4.57
7.33 + j2.08
0.00
0.00
0.00
0.00
0.13
0.23
0.05
0.05
0.05
0.05
0.25
0.15
+
+
+
+
+
+
+
+
+
+
+
+
j0.88
j1.12
j0.99
j1.23
j1.25
j2.30
j0.25
j0.25
j0.25
j0.25
j2.53
j1.42
From the output active and reactive power of each generator, pi jQi , and its
terminal voltage,Vgi0 ; i 1; 2; 3; 4 ; output current of each generator can be calculated as
Igi0
Pi jQi
;
Vgi0
i 1; 2; 3; 4
It gives
Ig10 7:0811 j1:1871
Ig20 7:7126 j1:2230
Ig30 7:0287 j2:1249
Ig40 6:2746 j4:1969
The imaginary voltage of each generator can be computed as
Q10 V
g10 jXq1Ig10 1:9976\70:7
E
Q20 V
g20 jXq2Ig20 2:2619\56:1
E
Q30 V
g30 jXq3Ig30 2:0896\40:0
E
Q40 V
g40 jXq4Ig40 2:3702\22:6
E
The phase of imaginary voltage is the angular position of each generator,
di0 ; i 1; 2; 3; 4. From di0 ; i 1; 2; 3; 4, computational results of output current and
terminal voltage of each generator can be transformed from the common xy
coordinate to the dq coordinate as (Eq. (5.49))
292
Table 6.5 Initial values of

state variables
idi
iqi
sin di
cos di
G1
G2
G3
G4
i0 (rad.)
i0
E0qi0
1.2341
1
1.0025
0.9788
1
1.0775
0.6990
1
1.0181
0.3942
1
1.1116
E0fdi0
2.0688
2.3456
2.1662
2.4601

vdi
vgxi
sin di
;
vgyi
vqi
cos di
cos di
sin di
cos di
sin di
vgxi
vgyi
Since E0fdi0 0, from the third equation of Eq. (5.24), it can have
Efdi0 Eqi0
E0q10 Vgq10 X0d1 Igd10 1:0025; Efd10 E0q10 Xd1 X0d1 Igd10 2:0688
Computational results of all the above are presented in Table 6.5. Initial values
of SVC are given as
Xsvcc 0:02; Xsvcl 0:01; Bsvc0 0
According to Eq. (6.34), it can have a0 90 .
6.3.1.2
Linearized HeffronPhillips Model
From Tables 6.1 and 6.3, the following network equations can be obtained:
where
Ig
0

Y
11
Y21
12
Y
22
Y
g
E
m
V
21
Y
22
Y
221
Y
222
Y
223
Y
0
j59:9880
0
0 7
7
7;
0
0
j59:9880
0 5
0
0
0
j59:9880
3
2
j59:9880
0
0 0 0
0
0
6
0
j59:9880 0 0 0
0
0 7
7
6
6
7;
4
0
0
0 0 0
0
j59:9880 5
0
0
0 0 0 j59:9880
0
3
2
j59:9880
0
0
0
6
0
j59:9880
0
0 7
7
6
7
6
7
6
0
0
0
0
7
6
7
6
6
0
0
0
0 7;
7
6
6
0
0
0
0 7
7
6
7
6
0
0
0
j59:9880 5
4
0
0
j59:9880
0

Y221 Y222 Y223 ,
3
2
3:9604 j99:5701
3:9604 j39:6040
0
6 3:9604 j39:6040 13:8614 j198:5712 9:90010 j99:0100 7
7
6
7
6
6
0
9:9010 j99:0100 23:6692 j118:0481 7
7
6
7
6
6
0
0
1:8002 j18:0020 7;
7
6
7
6
0
0
0
7
6
7
6
0
0
0
5
4
0
0
0
3
2
0
0
7
6
0
0
7
6
7
6
7
6 1:8002 j18:0020
0
7
6
7
6
6 3:6004 j35:6186 1:8002 j18:0020 7;
7
6
6 1:8002 j18:0020 34:5237 j118:1020 7
7
6
7
6
0
9:9001 j99:0100 5
4
0
0
3
2
0
0
7
6
0
0
7
6
7
6
7
6
0
0
7
6
7
6
6
0
0
7;
7
6
7
6 9:9010 j99:0100
0
7
6
7
6
4 13:8614 j198:5712 3:9604 j39:6040 5
3:9604 j39:6040 3:9604 j99:57010
6
11 6
Y
6
4
12
Y
j59:9880
293
By keeping only three nodes of generator terminals via deleting all the other
nodes in the above network admittance matrix, it can have
294
N Y
1 Y

11 Y
12 Y
Y
22 21
2
1:5150 j17:7819
6 1:1172 j15:2884
6
6
4 0:1867 j0:4071
0:5275 j0:9893
1:1172
5:0840
0:5275
1:4666
j15:2884
j21:9474
j0:9893
j 2:3934
0:1867
0:5275
1:9728
2:2208
0:5275
1:4666
2:2208
7:7304
j0:4071
j0:9893
j18:0523
j14:5440
3
j0:9893
j2:3934 7
7
7
j14:5440 5
j23:9732
Thus,

1
Y
1 jx0
Y
d
N
2
1:2870 9:3383i
6 1:4623 5:9214i
6
6
4 0:4321 0:4042i
0:6681 0:5845i
1:4623 5:9214i
0:4321 0:4042i
2:4883 11:1930i
0:6681 0:5845i
1:0309 0:8430i
0:6681 0:5845i
1:7113 98981i
2:0705 5:0608i
0:6681 0:5845i
1:0309 0:8430i 7
7
7
2:0705 5:0608i 5
3:3572 12:5138i

2
Fdd
Fqq
Gdd
6:9805
6 3:2377
6
6
4 0:4782
0:4958
2
3:2724
6 6:6150
6
6
4 0:4880
0:9808
2
9:3383
6 6:0988
6
6
4 0:5680
0:8876
Gqq
Hdd
5:4977
5:7984
0:8486
1:0508
3:9725
6:1187
0:5283
1:2240
5:3603
11:1930
0:7462
1:2719
0:6487
1:0716
6:8161
3:4812
0:1877
0:0321
3:6934
5:8082
0:1274
0:3772
9:8981
5:4489
3
0:8340
1:4891 7
7
7;
5:4893 5
5:0277
3
0:5123
0:4642 7
7
7;
2:6771 5
8:0130
3
0:1073
0:1341 7
7
7;
4:2062 5
12:5138
3
0:8811
1:3249 7
7
7;
3:4938 5
1:2870
6 0:0804
6
6
4 0:1656
2:9102
2:4883
0:5778
0:8035
0:4807
1:7113
0:0108
2
0:2002
6 0:0125
6
6
4 0:0258
0:3945 0:4563 3:3572

3
0:4527 0:0899 0:1371
0:3871 0:1250 0:2061 7
7
7;
0:0748 0:2662 0:5435 5
0:0017
1:4526
6 0:9487
6
Hqq 6
4 0:0884
0:1381
0:0614 0:0710 0:5222

3
0:8338
0:0198 0:0167
1:7411 0:0587
0:0209 7
7
7
0:1161 1:5397 0:6543 5
0:1979
0:8476
1:9466
295
Thus, according to Eq. (5.38), the following matrices are calculated:

2
5:9711
4:9200
0:7327 0:3184
6
6 4:2784
6
Fd 6
6 1:0557
4
1:0252
7
1:2632 0:8814 7
7
7;
6:1938 4:1129 7
5
0:9626
1:1694
4:2732
6:4053
0:5778
0:9782
0:1172
0:2833
6:4231
6
6 2:2267
6
Fq 6
6 0:4224
4
0:2615
0:5768
0:4099
9:5556
6
6 6:0083
6
Gd 6
6 0:4687
4
4:2382
0:2367
7
0:2801 7
7
7;
1:0323 0:3484 7
5
1:6646 2:6513
3
0:2888
0:9817
7
0:0148 0:7474 7
7
7;
10:3994 2:6752 7
5
0:8284
0:9169
5:1327
0:5850
6
6 0:1761
6
Gq 6
6 0:1130
4
0:0754
1:8849
11:8998
0:0617
1:6999
0:3921
1:5087
0:4498
0:4304
0:8213
13:6144
3
0:6178
7
0:7509 7
7
7;
1:8852 7
5
0:4386
0:4397
1:7610

K1 Vtd0 Iq0 X0d Fd Vtd0 Id0 Xq ;

K2 Iq0 Vtd0 Id0 Xq Gd Vtd0 Id0 Xq Gq ;

K3 I Xd X0d Gd ;

K4 Xd X0d Fd ;
K5 Vt01 Vq0 X0d Fd V1
t0 Vd0 Xq Fq ;
K6 Vt01 Vq0 Vt01 Vq0 X0d Gd V1
t0 Vd0 Xq Gq
Hence, coefcient matrices of the HeffronPhillips model of the example power
system as given by Eq. (5.43) are obtained as
296
4:4089
6 7:0963
6
K1 6
4 1:4589
1:8094
9:8813
6 2:4940
6
K2 6
4 0:0196
2
3
0:3923
0:2093 7
7
7;
2:9728 5
4:6359
7:3556
0:1653
0:4686
1:1060
1:5087
1:0858
12:2358
5:5378
5:9085 9:2266
3
0:9408
1:7693
1:0102
2:0583 7
7
7;
10:9008 2:4791 5
0:7701
0:6184 1:3160 13:3819
3
0:7064 0:0481
0:1636
2:9833 0:0025 0:1246 7
7
7;
0:0395 2:7332 0:4459 5
0:2018
2
2:5926
6 1:0014
6
K3 6
4 0:0781
0:1381
0:9952
6 0:7131
6
K4 6
4 0:1760
0:1528
0:8200
1:0705
0:8555
0:1221
0:2105
3:2691
3
0:0531
0:1469 7
7
7;
0:6855 5
0:1709 1:0323
0:1604 0:1949 0:7122 1:0675
2
3
0:0824
0:0069
0:0329 0:0425
6 0:1271 0:0315
0:0411 0:0546 7
6
7
K5 6
7;
4 0:0244 0:0056 0:0326 0:0626 5
0:0334 0:0103 0:0549 0:0986
2
3
0:5932 0:3146 0:0402 0:0499
6 0:1826 0:6569 0:0496 0:0618 7
6
7
K6 6
7:
4 0:0257 0:0583 0:5999 0:2974 5
0:0309
2
6:5 0
6 0 6:5
6
M6
4 0
0
0
0
2
50 0
6 0 50
6
KA 6
4 0
0
0
0
2
8 0 0
60 8 0
6
T0do 6
40 0 8
0 0 0
0:0712
0:1885
0:6390
3
2
0
0
0
7
6
0
0 7
60
7; D 6
5
40
6:175
0
0
6:175
0
2
3
0
0
0:1 0
6 0 0:1
0
0 7
6
7
7; T A 6
4 0
50 0 5
0
0 50
0
0
3
0
07
7
7
05
8
3
0
07
7
7;
05
0
3
0
0
0
0 7
7
7;
0:1 0 5
0 0:1
0
0
0
0
0
0
0
0
6.3.1.3
297
Linearized HeffronPhillips Model of the System with an SVC

Stabilizer to Be Installed
Consider that the SVC stabilizer is installed at node 9 in the example power system.
From Fig. 6.2 and Eq. (6.4), it can have
Ig
0
11
Y
21
Y
12
Y
22
Y
g
E
9
V
Thus,
Ig
1
1
12 Y
12 Y
21 E
g Y
g
21 Y
11 E
11 E
g Y
E g
Y
Y

y22 jbsvc
y22 jbsvc
where

11 Y
111
Y
2

112 ;
Y
j59:9880
7
7
7
9:5608 j92:1942 5
45:8239 j640:5978 10:0729 j231:8408
3
2
0
0
7
6
0
0
7
6
6
7;
4 1:2539 j20:7674 0:0170 j7:8663 5
6 8:9964 j89:9640
111 6
Y
6
4 33:5910 j254:3110
112
Y
0:0170 j7:8663
j59:9880
j59:9880
1:2218 j40:3316
3
6 8:9931 j149:9192 7
7
12 6
Y
6
7;
4 60:1463 j359:6609 5
2
96:3963 j907:8644
498:2591 j1004:9460
6 169:1906 j372:5732
21 6
Y
6
4 1:2703 j12:9860
1:2278 j32:6445

y22 757:0770 1400:3458
3T
7
7
7 ;
5
298

2
Fdd
5:5021
0:6541
6:0160
1:0512
0:8661
6:8227
0:6231
1:1592
3:9462
3:2904
6
6 6:2778
6
6
6 0:4221
4
3:9105
0:1630
5:8578
0:0239
0:4362
3:4441
0:8044
0:9453
4:9335
6:6832
9:3383
5:4253
0:1673
0:0150
11:1930
0:4145
6
6 3:4671
6
6
6 0:5079
4
2
Fqq
2
Gdd
Gqq
0:8800
7
1:4976 7
7
7;
5:4488 7
5
5:7286
3
0:4571
7
0:3961 7
7
7;
2:5858 7
5
0:1328
0:5381
0:8275
3:3572
0:2002
6 0:0091
6
6
4 0:0318
0:4335
0:3871
0:0883
0:1221
0:0801
0:2662
3
0:1381
0:2023 7
7
7;
0:4975 5
0:0207 0:0837 0:1287 0:5222

1:4526 0:8439
0:0260
0:0023
6 0:9487 1:7411 0:0645
0:0444
6
6
4 0:0864
0:1125 1:5397 0:6899
2
Hqq
0:1365
0:1895
0:8408
3
3
2
0:1043
0:0788
6 0:2946 7
6 0:1519 7
7
7
6
6
6
7; Lqq 6
7
4 0:6485 5
4 0:1730 5
1:6297
0:0638
2
Ldd
7
0:2855 7
7
7;
0:7229 9:8981 4:4351 7
5
0:8776 1:2181 5:4050 12:5138
3
2
1:2870 2:7871 0:5675 0:8875
7
6
6 0:0588 2:4883 0:7849 1:3007 7
7
6
6
7;
6 0:2047 0:5151 1:7113 3:1982 7
5
4
6
6 6:0990
6
6
6 0:5551
4
Hdd
7:0361
1:9466
3
7
7
7;
5

2
6:1526
6
6 4:4148
Fd 6
6 0:9953
4
1:0862
2
0:6169
6
6 2:0932
Fq 6
6 0:3695
4
4:9855 0:7232
299
0:4439
7
1:0000 7
7;
0:9652 6:4009 4:4404 7
5
1:2197 4:5385 6:8444
3
0:9820 0:0895
0:2756
7
1:7871
0:0404
0:2658 7
7;
0:2106 0:9843 0:4041 7
5
0:4845 0:3115 1:3867 2:1827
2
3
9:5940 4:3354 0:2644
0:8437
6
7
6 5:9571 11:9432 0:0136 0:5853 7
6
7;
Gd 6
7
4 0:4126 0:1860 10:4629 3:0182 5
0:7581 0:7453 4:9475 13:7236
2
3
0:6082 1:6579 0:3971 0:6468
6
7
6 0:2375 1:5002 0:4534 0:7684 7
6
7;
Gq 6
7
4 0:1473 0:4578 0:8635 1:7847 5
0:1306 0:4865 0:5748 1:7280
3
2
2
3
0:0604
0:1662
7
6
6
7
6 0:0793 7
6 0:3477 7
7
6
7
Ld 6
7
6 0:7077 7; Lq 6
7
6
4
5
4 0:0886 5
1:6776
0:0548
6:6100
1:1952
Substituting into Eqs. (5.40) and (5.41), it can have

K1 Vtd0 Iq0 X0d Fd Vtq0 Id0 Xq Fq ;

K2 Iq0 Vtd0 Iq0 X0d Gd Vtq0 Id0 Xq Gq ;

Kp Vtd0 Iqo X0d Ld Vtd0 Ido X0q Lq ;

K3 I Xd X0d Gd ;

K4 Xd X0d Fd ;

Kq Xd X0d Ld ;
0
1
K5 V1
t0 Vq0 Xd Fd Vt0 Vd0 Xq Fq ;
1
0
1
K6 V1
t0 Vq0 Vt0 Vq0 Xd Gd Vt0 Vd0 Xq Gq ;
0
1
Kv V1
t0 Vq0 Xd Ld Vt0 Vd0 Xq Lq
300
4:6390
4:6990
0:2294
0:2894
7
0:0586 7
7
7;
5:6138 3:3503 7
5
1:5697 1:2865 5:2531 8:1092
3
2
10:1342 0:8111
0:9219
1:7323
7
6
6 2:5271 12:5659 0:9623
1:9405 7
7
6
K2 6
7;
6 0:0495
0:7864 11:4295 1:7433 7
5
4
0:1357 0:6002 1:4532 14:2532
3
2
2:5990 0:7226 0:0441
0:1406
7
6
6 0:9929 2:9905 0:0023 0:0976 7
7
6
K3 6
7;
6 0:0688 0:0310 2:7438 0:5030 7
5
4
6
6 6:7233 7:1872
6
K1 6
6 1:2956 0:9680
4
0:5225
0:1263 0:1242
0:8246
3:2873
0:8309
0:1205
0:0740
1:0254
0:0776 0:0029
6
6 0:1258 0:0326
6
K5 6
6 0:0231 0:0035
4
0:0300
7
0:1667 7
7
7;
0:7401 7
5
1:1407
3
0:0454
7
0:0571 7
7
7;
0:0566 7
5
0:0306 0:0063
0:0511
0:0881
6
6 0:7358 1:1017
6
K4 6
6 0:1659 0:1609
4
2
0:1810 0:2033
0:5876
6 0:1844
6
K6 6
4 0:0290
0:3150
0:0428
0:6494
0:0633
0:0528
0:5884
0:1992
1:0668
0:7564
0:0293
0:0362
0:0593
0:0735 7
7
7;
0:3030 5
0:0352 0:0778 0:1962 0:6259

3
3
3
2
2
0:2106
0:0277
0:0033
7
7
7
6
6
6
6 0:3466 7
6 0:0580 7
6 0:0000 7
7
7
7
6
6
6
KP 6
7; Kq 6
7; Kv 6
7:
6 0:5811 7
6 0:1179 7
6 0:0059 7
5
5
5
4
4
4
1:0042
0:2796
0:0364
2
6.3.2
Selection of Installing Locations of Stabilizers
6.3.2.1
Selection of the PSS Installing Locations
301
From the linearized HeffronPhillips model of the N-machine power system,

eigenvalues of the state matrix are calculated as
1;2 5:1604 j6:0402
k
3;4 5:3163 j4:9927
k
5;6 0:1979 j6:1696
k
7;8 0:1707 j6:3459
k
9;10 0:1641 j3:2658
k
11;12 5:8428 j0:0805
k
13;14 4:1998 j0:0862
k
15;16 0 j0
k
According to Eqs. (5.60) and (5.63), the correlation ratio of electromechanical
loop can be calculated to identify the electromechanical oscillation modes of the
example power system as
r1;2 0:0080
r3;4 0:0503
9
r5;6 23:5268 >
=
r7;8 27:3818
>
;
r9;10 17:0246
electromechanical modes
r11;12 0:1747
r13;14 0:1658
Hence, three electromechanical oscillation modes of the example power system
are
5;6 0:1979 j6:1696
k
7;8 0:1707 j6:3459
k
9;10 0:1641 j3:2658
k
302
9;10 is 0.0437. Right and left

Damp ratio of inter-area oscillation mode k

eigenvectors corresponding to k9 0:1641 j3:2658 are calculated as
3
3
2
2
0:0352 j0:0797
0:7994 j0:5947
6 0:7288 j0:4218 7
6 0:0312 j0:0629 7
7
7
6
6
6 0:7969 j0:6005 7
6 0:1850 j0:0850 7
7
7
6
6
6 0:7313 j0:4160 7
6 0:1683 j0:0753 7
7
7
6
6
6 73:0984 j88:6036 7
6 0:0007 j0:0003 7
7
7
6
6
6 52:7805 j81:4813 7
6 0:0006 j0:0002 7
7
7
6
6
6 73:7524 j88:2869 7
6 0:0008 j0:0016 7
7
7
6
6
6 52:1265 j81:7980 7
6 0:0007 j0:0014 7
7
7; w
6

v9 6
9
6 0:5131 j1:5515 7
6 0:0087 j0:0165 7
7
7
6
6
6 0:3709 j1:8021 7
6 0:0050 j0:0249 7
7
7
6
6
6 0:5572 j1:6701 7
6 0:0052 j0:0066 7
7
7
6
6
6 0:4024 j1:9651 7
6 0:0043 j0:0093 7
7
7
6
6
6 0:0002 j0:0191 7
6 0:5060 j0:1325 7
7
7
6
6
7
6 0:0023 j0:0214 7
6
0:7359
7
7
6
6
4 0:0002 j0:0206 5
4 0:1030 j0:1762 5
0:0026 j0:0233
0:1964 j0:1248
1. Selection of installing locations based on the participation factor

From Eq. (6.140), the participation factor of each generator can be calculated to
examine how much each generator is involved in the oscillation modes. The results
are given in Table 6.6.
9 0:1641 j3:2658, it can be seen that generator 3 is the best installing
For k
location.
2. Selection of installing locations by damping torque analysis
The linearized HeffronPhillips model for the PSSs to be installed is given in the
above section.
sDd xo IDx
sDE0 T01 K3 DE0 K4 Dd DE0
q
d0
fd
0
1
0
sDE0fd T1
A DEfd TA KA K5 Dd K6 DEq Fk Dupssk
Table 6.6 Participation factor of each generator

Generator
1
2
3
4
Electromechanical oscillation modes

0.1979 j6.1696
0.1707 j6.3459
0.1641 j3.2658
0.2269
0.2843
0.0017
0.0010
0.0868
0.0591
0.2032
0.1551
0.0000
0.0031
0.2053
0.3062
303
From Eq. (6.145), the DTA can be used to select installing location of the PSSs.
If the PSS is installed in generator 3, F3 0 0 1 0 T . From Eq. (6.115), it
9 0:1641 j3:2658.
can have with s k
9 DTPSS M1 K2 I sTA K3 sTd0 KA K6 1 KA F3
FPSS k
Dupss3
2
3
0:0365 j0:0099
6 0:0401 j0:0102 7
6
7
6
7
4 0:1147 j0:0109 5
0:1129 j0:0107
9 , the elements of right eigenvector corresponding to Dxj ; j 1; 2; 3; 4 are
For k
8
v95
>
>
<
v96
v
>
>
: 97
v98
0:0007 j0:0003
0:0006 j0:0002
0:0008 j0:0016
0:0007 j0:0014
Let the feedback signal be Dx3 , and thus,

Dx3 Dy cT3 X
where the 7th element of C3 is 1 and all other elements are zero. From Eq. (6.154),
it can have
8
cT3 v9
>
>
9
c1 k
>
>
>
v95
>
>
>
T
>

>
>
9 c3 v9
>
< c2 k
v96
T
>

c
>
9 3 v9
>
c3 k
>
>
>
v97
>
>
>
T
>

c
>
>
9 3 v9
: c4 k
v98
v97
v95
v97
v96
v97
v97
v97
v98
1:7522 j1:5478
2:2552 j1:8217
1
1:1042 j0:0115
From Eq. (6.154), sensitive index Sij can be calculated as

8
291 v291 0:0755
S91 w
>
>
>
<S w
292 v292 0:0493
92
>
293 v293 0:1985
S
w
93
>
>
:
294 v294 0:1544
S94 w
j0:0427
j0:0327
j0:0433
j0:0150
6:187
304
6
Generator 1
Ddamp1 = 0.0092 j0.0074
S91 = 0.0755 j0.0427
Generator 2
S92 = 0.0493 j0.0327
PSS
Ddamp2 = 0.0124 j0.0144

Generator 3
Ddamp3 = 0.0881 j0.0931
9,10
S93 = 0.1985 + j0.0433
Generator 4
S94 = 0.1544 + j0.0150
Ddamp4 = 0.0107 + j0.0526
Fig. 6.18 The PSS to be installed on G3 to affect the inter-area oscillation mode
According to Fig. 6.14 and Eq. (6.145), the DTA can be calculated as
DTA3
4
X

9 c k
9 S
F
9j 0:0576
pssj k
j
j1
where
2
3
0:0092 j0:0074
6 0:0124 j0:0144 7
6
7
9 c k

pssj k
F
7
j 9 6
4 0:0881 j0:0931 5
0:0107 j0:0526
Figure 6.18 shows how the PSS to be installed on generator 3 will affect the
inter-area oscillation mode. When the PSS is installed on generators 1, 2, or 4, the
DTA can be calculated similarly as
DTA1 0:0150; DTA2 0:0114; DTA4 0:0354
Obviously, generator 3 is the best installing location.
6.3.2.2
Selection of Installing Location of the SVC Stabilizer
From the HeffronPhillips model of the example power system installed with the
SVC, eigenvalues of the state matrix are calculated as
305
1;2 0:1960 j6:1676 )

k
electromechanical modes
3;4 0:1607 j6:3093
k
5;6 5:5111 j5:6757
k
7;8 0:2433 j3:1216 electromechanical mode
k
9;10 5:9764 j4:4089
k
11 5:8189
k
12 5:5025
k
13 4:9368
k
14 4:1875
k
15 2:4164
k
16 0
k
17 0
k
An SVC stabilizer is to be installed to improve the damping of the inter-area
7 0:2433 j3:1216. Right and left eigenvectors corresponding
oscillation mode k

to k7 0:2433 j3:1216 are calculated as
3
3
2
0:0624 j0:0911
0:7677 j0:5166
6 0:0563 j0:0725 7
6 0:6952 j0:3532 7
7
7
6
6
6 0:1749 j0:0796 7
6 0:7436 j0:4640 7
7
7
6
6
6 0:1611 j0:0693 7
6 0:7193 j0:4058 7
7
7
6
6
6 0:0008 j0:0005 7
6 69:2025 j87:3225 7
7
7
6
6
6 0:0006 j0:0004 7
6 48:8981 j80:1477 7
7
7
6
6
6 0:0008 j0:0014 7
6 62:6622 j84:9249 7
7
7
6
6
6 0:0007 0:0013i 7
6 55:4385 j82:5452 7
7
7
6
6
7 7 6 0:5691 j1:5609 7
7 6
v
7
6 0:0097 j0:0170 7; w
6
6 0:0064 j0:0265 7
6 0:4629 j1:9346 7
7
7
6
6
6 0:0029 j0:0014 7
6 0:1767 j0:8646 7
7
7
6
6
6 0:0032 j0:0021 7
6 0:1625 j0:8590 7
7
7
6
6
6 0:5220 j0:1297 7
6 0:0010 j0:0194 7
7
7
6
6
6 0:7683 j0:0000 7
6 0:0014 j0:0232 7
7
7
6
6
6 0:0338 j0:1256 7
6 0:0010 j0:0103 7
7
7
6
6
4 0:0012 j0:1189 5
4 0:0011 j0:0101 5
0:0005 j0:0008
3:5217 j14:0586
2

Davi Tvsvci sDVi

2ai0
Dbsvci 1cos
pxsvcli Davi Dasi
306
Let the transfer function of the SVC voltage controller be

Tvsvci s Kvp
Kvi
s
i
where Kvp 1; Kvi 8. Let XVSVCi DV
s . It can have

X_ vsvci DVi
Davi Kvi Xvsvci Kvp DVi

h
i
8
viy0
v
>
< Avsvci 0; Bvsvci Vix0i0 Vi0
h
i
>
: cvsvci Kvi ; Dvsvci vix0 viy0
Vi0 Vi0
Hence, from Eq. (6.48), it can be obtained that
bssvci

1 cos 2ai0 vyi0
26:0361
vxi0
49:5981
pxsvcli
Substituting the above into Eq. (6.103), it can have

3
0
7
6 0
7
6
7
6 0
7
6
7
6 0
7
6
6 0:0215 7
7
6
6 0:0284 7
7
6
7
6
2
3 6 0:0818 7
6 0:0905 7
0
7
6
7
B 4 B2k 5 6
6 0:0520 7
6 0:0627 7
B3k
7
6
6 0:0882 7
7
6
6 0:0372 7
7
6
6 3:0862 7
7
6
6 4:3028 7
7
6
6 143:3026 7
7
6
4 178:5977 5
2
0:5828
The feedback signal of the SVC stabilizer is taken to be the active power, Pij ,
delivered through the node where the SVC is to be installed. Linearization of
Vi Vj
Vi Vj
Pij X
sinui uj X
sinuij gives
DPij
307
Vi Vj
Vi0 Vj0
Vi0 Vj0
sinuij0 DVi
sinuij0 DVj
cosuij0 Dui
X
X
X
Vi0 Vj0
cosuij0 Duj

X
where X is the line reactance between nodes i and j. In the common xy coordinate,
DVk and Duk ; k i; j can be written as
Vyk0
Vxk0
DVxk
DVyk
Vk0
Vk0
Vyk0
Vxk0
Duk
DVx 2 DVyk
2
Vk0
Vk0
DVk
Hence,
DPij
Vi0 Vj0
Vi0 Vj0
Vyi0
Vxi0
sin uij0
sin uij0
DVxi
DVyi
X
Vi0
X
Vi0
Vi0 Vj0
Vxj0
Vi0 Vj0
Vyj0
sin uij0
sin uij0
DVxj
DVyj
X
Vj0
X
Vj0
Vi0 Vj0
Vyi0
Vi0 Vj0
Vxi0
cos uij0 2 DVxi
cos uij0 2 DVyi
X
X
Vi0
Vi0
Vi0 Vj0
Vyj0
Vi0 Vj0
Vxj0
cos uij0
cos uij0 2 DVyj

DVxj
X
X
V2j0
Vj0
which can be written as

DPij a11 DVxi a12 DVyi a21 DVxj a22 DVyj
where
8
Vi0 Vj0
Vyi0
Vxi0 Vi0 Vj0
>
a11
sin uij0
cos uij0 2

>
>
>
X
Vi0
X
Vi0
>
>
>
>
>
V
V
V
V
V
V
i0 j0
yi0
i0 j0
xi0
>
>
>
< a12 X sin uij0 Vi0 X cos uij0 V2
i0
>
>
a21
>
>
>
>
>
>
>
>
Vi0 Vj0
Vyj0 Vi0 Vj0
Vxj0
>
>
sin uij0
cos uij0 2

: a22
X
Vj0
X
Vj0
Vi0 Vj0
Vxj0 Vi0 Vj0
Vyj0
sin uij0
cos uij0 2
X
Vj0
X
Vj0
6:188
308

DV Y1
M Cxy DX
Hence, the voltage at any node can always be reconstructed by state variables,
that is
DVxi YixT DX
DVyi YiyT DX
6:189
DVxj YjxT DX
DVyj YjyT DX
2
3
Dd
Because Dy DPij cTk 4 Dx 5 cTk DX; from Eqs. (6.188) and (6.189) above,
DZ
it can have
ck a11 Yix a12 Yiy a21 Yjx a22 Yjy
6:190
For the example power system, when the SVC stabilizer is to be installed at node
9, the feedback signal is taken to be P89, the active power delivered through node 8
to node 9. From the results of power flow calculation, it can have
V80 0:8636; V90 0:8799; u80 0:2001; u90 0:4834
Thus,
a11 0:3999; a12 0:8355; a21 0:2098; a22 0:8856
According to Eq. (6.101), for the example power system, it has

YM1 YM1
2
Y1
M1
7:8484
6 24:9097
6
6 0:9779
6
6 15:1586
6
6
0
6
6
0
6
6
0
6
6
0
6
4 0:1448
2:0653
46:0583
4:9233
15:1586
0:9779
0
0
0
0
2:0653
0:1448
1
0:9779
15:1586
15:5048
33:8809
0
0
0
0
0:6714
5:1397
YM1

2
15:1586
0:9779
45:9152
6:0720
0
0
0
0
5:1397
0:6714
0
0
0
0
13:3470
40:9433
0:0156
7:8445
1:2690
12:9499
3
0
7
0
7
7
0
7
7
0
7
35:9030 7
7
10:8393 7
7
7:8445 7
7
0:0156 7
7
12:9499 5
1:2690
0
6
0
6
6
0
6
6
0
6
6 0:0156
6
6 7:8445
6
6 10:1321
6
6 66:5814
6
4 1:2300
32:5540
1
YM2
309
0
0:1448
0
2:0653
0
0:6714
0
5:1397
7:8445 1:2690
0:0156
12:9499
49:4841
1:2300
7:7021
32:5540
32:5540
3:9996
1:2300 98:4466
3
2:0653
0:1448 7
7
5:1397 7
7
0:6714 7
7
12:9499 7
7
1:2690 7
7
32:5540 7
7
1:2300 7
7
65:4289 5
48:9889

Cxy Cxy1
2
Cxy1
18:5288
6 5:8704
6
6
0
6
6
0
6
6
0
6
6
0
6
6
0
6
6
0
6
4
0
0
Cxy2
0
0
0
0
0
0
0
0
0
0
28:8999
8:0494
0
0
0
0
0
0
0
0
0
6
0
6
6
0
6
6
0
6
6 18:9260
6
6
6 23:2767
6
0
6
6
0
6
4
0
0
0
0
0
0
0
0
0
0
0
0
Cxy2
0
0
0
0
0 18:4483
0 3:2065
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
16:9781
7:6725
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
25:8688
15:1922
0
0
0
0
0
0
0
0
0
0
0
0
11:7722
27:5938
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
18:5570
5:6022
0
0
0
0
3
0
7
0
7
7
0
7
7
0
7
7
0
7
7
0
7
7
0
7
7
0
7
201:7248 5
398:2669
Thus,
1
Y1
M Cxy YM Cxy 1
Y1
M Cxy 2
Y1
M Cxy 3
3
0
07
7
07
7
07
7
07
7
07
7
07
7
07
7
05
0
310
Y1
M Cxy 1
0:4930
6 0:5948
6
6 0:2527
6
6 0:3017
6
6 0:0098
6
6 0:0505
6
6 0:0006
6
6 0:0664
6
6 0:0201
6
6 0:0878
6
6 0:3899
6
6 0:5133
6
6 0:2474
6
6 0:3753
6
6 0:1828
6
6 0:3638
6
6 0:0820
6
6 0:2283
6
6 0:0090
6
6 0:0763
6
4 0:0030
0:0613
Y1
M Cxy2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0:0341
6 0:0111
6
6 0:0378
6
6 0:0184
6
6 0:0050
6
6 0:6827
6
6 0:0268
6
6 0:2835
6
6 0:1330
6
6 0:2698
6
6 0:0393
6
6 0:0200
6
6 0:0484
6
6 0:0325
6
6 0:0593
6
6 0:0453
6
6 0:0973
6
6 0:1591
6
6 0:0806
6
6 0:3305
6
4 0:0376
0:5445
0:2838
0:6333
0:1235
0:1663
0:0107
0:0233
0:0192
0:0341
0:0466
0:0505
0:2505
0:4797
0:1774
0:2541
0:1857
0:2132
0:1175
0:1331
0:0326
0:0422
0:0190
0:0314
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0 0:1810
0 0:2330
0 0:2656
0 0:5832
0 0:0048
0 0:0571
0 0:0074
0 0:0765
0 0:0382
0 0:1034
0 0:1775
0 0:3233
0 0:1587
0 0:4580
0 0:0946
0 0:4303
0 0:0282
0 0:2697
0 0:0220
0 0:0892
0 0:0051
0 0:0706
0:0519
0:0061
0:0564
0:0075
0:5385
0:4218
0:1369
0:1461
0:1905
0:1561
0:0574
0:0035
0:0660
0:0003
0:0756
0:0047
0:1345
0:0811
0:2184
0:1886
0:4166
0:3213
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0:3392
0:0626
0:5422
0:3967
0:0249
0:0103
0:0343
0:0197
0:0701
0:0355
0:3858
0:1282
0:4654
0:2195
0:4404
0:1542
0:2581
0:0956
0:0522
0:0278
0:0354
0:0179
0:0538
0:0081
0:0591
0:0145
0:0196
0:2397
0:1462
0:5675
0:2048
0:3211
0:0610
0:0176
0:0732
0:0309
0:0875
0:0451
0:1479
0:1849
0:1665
0:3922
0:0745
0:3041
3
0
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
05
0
3
0
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
07
7
05
0
Y1
M Cxy3
0:0701
6 0:0180
6
6 0:0756
6
6 0:0251
6
6 0:1712
6
6 0:0275
6
6 0:5319
6
6 0:2074
6
6 0:2520
6
6 0:1031
6
6 0:0764
6
6 0:0190
6
6 0:0857
6
6 0:0214
6
6 0:0957
6
6 0:0208
6
6 0:1758
6
6 0:0414
6
6 0:3350
6
6 0:1216
6
4 0:2343
0:0690
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
311
3
0:5669
0:8487 7
7
0:5768 7
7
1:2875 7
7
2:2984 7
7
4:5241 7
7
1:8674 7
7
5:6191 7
7
1:6762 7
7
6:9322 7
7
0:5434 7
7
1:1876 7
7
0:4571 7
7
1:6917 7
7
0:3264 7
7
2:1281 7
7
1:0071 7
7
4:5797 7
7
1:8476 7
7
6:1937 7
7
2:1590 5
5:2016
Thus,
2
Y8x
Y8y
Y9x
0:0650
6 0
6
6 0:1252
6
6 0
6
6 0:0082
6
6 0
6
6 0:2609
6
6 0
6
Y9y 6
6 0:0972
6 0
6
6 0:1353
6
6 0
6
6 0:1476
6
6 0
6
6 0:1755
6
4 0
0:9722
0:2336
0
0:1235
0
0:2696
0
0:0767
0
0:1612
0
0:0838
0
0:1818
0
0:0396
0
4:6129
0:0246
0
0:0471
0
0:0427
0
0:0685
0
0:1345
0
0:1935
0
0:2072
0
0:2542
0
1:6229
3
0:0886
0 7
7
0:0467 7
7
0 7
7
0:1022 7
7
0 7
7
0:0288 7
7
0 7
7
0:2724 7
7
0 7
7
0:1609 7
7
0 7
7
0:3158 7
7
0 7
7
0:1013 7
7
0 5
6:9396
312
Finally, according to Eq. (6.190), it is obtained that

3
0:0856
6
0 7
7
6
6 0:1020 7
7
6
6
0 7
7
6
6 0:1225 7
7
6
6
0 7
7
6
6 0:1285 7
7
6
6
0 7
7
6
7
ck 6
6 0:0959 7
6
0 7
7
6
6 0:0589 7
7
6
6
0 7
7
6
6 0:1122 7
7
6
6
0 7
7
6
6 0:0398 7
7
6
4
0 5
2:3398
2
1. Selection of installing location by using damp torque analysis

2
Bk B2k A23 sI A33 1 B3k
0:0740
6 0:0102
6
4 0:0205
0:0206
3
j0:0070
j0:0092 7
7
j0:0287 5
j0:0374
From Eq. (6.154), it can be calculated that

8
cTk v7
>
>
7
c1 k
>
>
>
v75
>
>
>
T
>

>
c
>
7 k v7
>
< c2 k
v76
T
>

c
>
7 k v7
>

k
c
>
>
> 3
v77
>
>
>
T
>

c
>
>
7 k v7
: c4 k
v78
25:8592 j63:8552
26:4325 j78:5172
6:2615 j39:1056
7:7533 j42:6894
313
Thus, the following coefcients in Eq. (6.147) are obtained:

3
0:0500 j0:0515
6 0:0474 j0:0663 7
i c k

7
6
kj k
B
j i 4
0:1877 j0:1421 5
0:2250 j0:1462
2
From Eq. (6.142), the sensitive index Sij can be calculated as

8
T291 v291 0:0950 j0:0377
S91 w
>
>
>
<
T291 v292 0:0648 j0:0305
S92 w
> S93 w
T291 v293 0:1670 j0:0220
>
>
:
T291 v294 0:1440 j0:0156
S94 w
DTA9 0:3585
Figure 6.19 illustrates how the SVC stabilizer to be installed at node 9 contributes the damping to the inter-area oscillation mode.
When the installing location of SVC stabilizer is at node 7 and node 8, the
following results can be obtained similarly:
DTA7 0:2316; DTA8 0:0478
Obviously, the best installing location for the SVC stabilizer is at node 9.
Generator 1
S 71 = 0.0950 j0.0377
Ddamp1 = 0.0500 j0.0515
Generator 2
SVC
Stabilizer
S 72 = 0.0648 j0.0305
Ddamp2 = 0.0474 j0.0663
Generator 3
Ddamp3 = 0.1877 + j0.1421
7,8
S 73 = 0.1670 + j0.0220
Generator 4
Ddamp4 = 0.2250 + j0.1462
S 74 = 0.1440 + j0.0156
Fig. 6.19 The SVC stabilizer to be installed at node 9 contributing damping to the inter-area
oscillation mode
314
2. Selection of installing location based on the residue index

9 and W
T into Eq. (6.149), it can have
Substituting Bk ; ck ; V
9
ik
R
bikcik 5:5927 j0:94650:0498 j0:0389 0:3153 j0:1705
ik j 0:3585, being equal to DTA9 0:3585.
thus, jR
Residue index of node 7 and node 8 is calculated as
Node 7:
ikcik 1:3986 j0:89580:0854 j0:1103 0:2182 0:0777i
ik b
R

jRik j 0:2316
Node 8:
ik
R
bikcik 0:4780 j0:15530:0791 j0:0528 0:0296 j0:0375

jRik j 0:0478
They are equal to DTA7 0:2316 and DTA8 0:0478, respectively.
6.3.2.3
Robust Installing Locations and Feedback Signals of the SVC

Stabilizer
Let six candidate installing locations and feedback signals of the SVC stabilizer in
the example power system be
8
u:
>
>
> 1
>
u
:
>
>
< 2
u3 :
u4 :
>
>
>
>
u:
>
>
: 5
u6 :
node
node
node
node
node
node
7
7
8
8
9
9
DP67
DP78
DP78
DP89
DP89
DP910
Consider Xl the set of operating conditions of the example power system as

given in Table 6.7.
Obviously, in order to apply the criteria of Eqs. (6.162) and (6.163), the effectiveness of the stabilizer has to be examined at ten given system operating conditions. Table 6.8 presents the results of residue calculation.

Table 6.7 System operating
conditions
315
MW
Load at node 7
Load at node 9
l1
l2
l3
l4
l5
l6
l7
l8
l9
l10
927 + j100
937 + j100
947 + j100
957 + j100
967 + j100
977 + j100
987 + j100
997 + j100
1007 + j100
1017 + j100
1767
1767
1767
1767
1767
1767
1767
1767
1767
1767
Table 6.8 Residue (Cu; l)

l1
l2
l3
l4
l5
l6
l7
l8
l9
l10
+
+
+
+
+
+
+
+
+
+
j100
j100
j100
j100
j100
j100
j100
j100
j100
j100
u1
u2
u3
u4
u5
u6
0.2263
0.2283
0.2302
0.2319
0.2336
0.2353
0.2369
0.2384
0.2399
0.2413
0.2240
0.2260
0.2279
0.2298
0.2316
0.2334
0.2351
0.2368
0.2384
0.2400
0.0585
0.0559
0.0532
0.0505
0.0478
0.0451
0.0424
0.0397
0.0371
0.0344
0.0534
0.0512
0.0489
0.0465
0.0442
0.0418
0.0394
0.0370
0.0346
0.0322
0.3485
0.3509
0.3534
0.3559
0.3585
0.3610
0.3637
0.3663
0.3689
0.3716
0.0311
0.0314
0.0317
0.0319
0.0322
0.0325
0.0327
0.0330
0.0333
0.0335
The following two indices are used to examine the effectiveness of the stabilizer.
8
1: min Cu; l
>
>
l
<
max Cu; l min Cu; l
l
l
>
>
: 2:

Cu; l

where Cu;
l is the mean value of the residue (Cu; l) at all operating conditions. Two indices above are applied for u1 u6 , and the results of selection are
given in Table 6.9.
In order to consider the robustness of the stabilizer to the variations of power
system operating conditions, the criteria (6.161), (6.162), and (6.163) should be
taken into account. That is
316
Table 6.9 Results of indices

for u1 u6
u1
u2
u3
u4
u5
u6
Index 1
0.2263
0.2240
0.0478
0.0442
0.3485
0.0311
Index 2
0.0640
0.0689
0.5187
0.4939
0.0642
0.0742
8
< Maxmin C/; l
/
: Minmax C/; l min C/; l

l
Obviously, from Table 6.9, the nal selection should be u5 . That is, the
installing location is at node 9, and the feedback signal is DP89 .
6.3.2.4
Design of Robust SVC Stabilizer
According to the results obtained in the previous section, the SVC stabilizer is
installed at node 9 and the integral of P89 is used as the feedback signal. The
stabilizer is designed at operating condition l1 , where it is estimated to be least
effective. At l1 , the inter-area oscillation mode is k7 0:2473 j3:0340. Right and
7 0:2473 j3:0340 are calculated as
left eigenvectors corresponding to k
3
3
2
0:0512 j0:0923
0:7809 j0:5431
6 0:0466 j0:0745 7
6 0:7129 j0:3833 7
7
7
6
6
6 0:1739 j0:0788 7
6 0:7625 j0:4873 7
7
7
6
6
6 0:1611 j0:0682 7
6 0:7314 j0:4390 7
7
7
6
6
6 0:0008 j0:0003 7
6 75:8300 j91:8156 7
7
7
6
6
6 0:0006 j0:0003 7
6 55:1927 j84:9901 7
7
7
6
6
6 0:0007 j0:0013 7
6 68:6860 j90:0933 7
7
7
6
6
6 0:0006 j0:0012 7
6 62:3367 j86:7124 7
7
7
6
6

7
6
9 6 0:0102 j0:0178 7; W
V
7 9 6 0:6701 j1:6183 7
6
6 0:0069 j0:0278 7
6 0:5915 j2:0830 7
7
7
6
6
6 0:0033 j0:0024 7
6 0:2084 j0:8511 7
7
7
6
6
6 0:0038 j0:0026 7
6 0:2230 j0:8620 7
7
7
6
6
6 0:5193 j0:1312 7
6 0:0022 j0:0204 7
7
7
6
6
6 0:7712 j0:0000 7
6 0:0002 j0:0253 7
7
7
6
6
6 0:0006 j0:1244 7
6 0:0005 j0:0102 7
7
7
6
6
4 0:0208 j0:1235 5
4 0:0003 j0:0104 5
0:0006 j0:0010
4:2900 j15:2077
2
Taking the same procedure of computation, it can have
317
3
3T
2
0
0:0773
7
7
6
6 0
0
7
7
6
6
7
6
6 0:1042 7
0
7
7
6
6
7
7
6
6 0
0
7
7
6
6
6 0:0199 7
6 0:1159 7
7
7
6
6
7
6 0:0270 7
6 0
7
7
6
6
6 0:0793 7
6 0:1367 7
7
7
6
6
7
6 0:0914 7
6 0
7
7
6
6
7; Ck 6 0:0881 7
0:0520
Bk 6
7
7
6
6
7
6 0:0643 7
6 0
7
7
6
6
6 0:0821 7
6 0:0488 7
7
7
6
6
7
6 0:0384 7
6 0
7
7
6
6
6 2:5208 7
6 0:1051 7
7
7
6
6
7
6 2:6311 7
6 0
7
7
6
6
6 143:597 7
6 0:0307 7
7
7
6
6
5
4 176:7661 5
4 0
0:5767
2:3522
2
9 and W
T into Eq. (6.149), the residue is calculated as
Substituting Bk Ck V
9
ik
R
bikcik 5:8380 j1:13650:0448 j0:0378 0:3485\150:8661
Let the transfer function of the SVC stabilizer be
Tsvc s = Ksvc
sTw
1 1 + sT2 1 + sT4
= Ksvc Tk s
1 + sTw 1 + sT 1 + sT1 1 + sT3
where T1 0:79s; T3 0:79s; T 0:01s; Tw 10s.

Firstly, T2 and T4 are set such that the SVC stabilizer to compensate the phase of
the residue index to ensure the SVC stabilizer to move the oscillation mode horizontally towards the left on the complex plane. The result is
T2 0:2371s,T4 0:2703s
It gives

7 1\ 150:8661
Tk k
To determine the gain value of the SVC stabilizer, Ksvc , a direct searching
7 to k
by minimizing the
method in nonlinear programming can be used to move k
7
following objective function:
318

7 K) Re k
K
f K Re k
7

K 0:45. The steps of direct searching are as follows:
where Re k
7
Step 1: Set the initial searching scale S 2 and minimum searching scale
Sm 0:0001. Choose K0 50 as the starting point of the searching. The oscilla7 0:2031 j4:6030. Thus, the initial value of
tion mode is calculated as k
objective function is f0 0:0610.

Step 2: Search from K0 to K0 S, if f K0 S \f K0 let K0 ( K0 S and

execute Step 2 again; if not, search from K0 to K0 S, if f K0 S \f K0 let
K0 ( K0 S and execute Step 2 again; if not, Step 2 fails and go to Step 3.
Step 3: If searching in Step 2 fails, reduce the searching scale S by half and then go
back to Step
2.
Step 4: If f K0 \Sm , stop searching and take K0 as the solution K .
For
theSVC stabilizer, in 36 steps, the searching stops to have K = 112.5000
with f K0 0. Thus, the design of the SVC stabilizer is completed. Parameters of
SVC stabilizer are set to be
Ksvc 112:5000; T2 0:2371 s; T4 0:2704 s
With the SVC stabilizer installed, eigenvalues of the state matrix are calculated as
1;2 49:7492 j85:8280
k
)
3;4 0:1615 j6:3364
k
5;6 0:1840 j6:1485
k
7;8 5:7637 j5:5700
k
9;10 0:4500 j4:5671
k
11;12 4:3308 j3:9562
k
13;14 0:2099 j22:0032
k

k15 6:3962

k16 3:4632
17;18 4:0904 j0:2451
k
19 4:3536
k
20 5:6946
k
21 0
k
22 0
k
electromechanical mode
electromechanical mode

Fig. 6.20 Simulation result
with and without the SVC
stabilizer
319
1 3 /degree
t/s
Table 6.10 Oscillation mode

when the SVC stabilizer is
installed and the system is
operated at varied operating
conditions
Operating condition
Eigenvalue
l1
l2
l3
l4
l5
l6
l7
l8
l9
l10
0.4500
0.4510
0.4517
0.4526
0.4576
0.4633
0.4706
0.4838
0.4870
0.4901
Damp ratio
j4.5671
j4.5600
j4.5553
j4.5512
j4.5344
j4.5269
j4.5197
j4.4630
j4.4068
j4.3511
0.0983
0.0984
0.0987
0.0990
0.1004
0.1018
0.1036
0.1078
0.1098
0.1119
It can be observed that the mode of interest is moved to 0:4500 j4:5671

which is well damped.
The simulation result of the example power system at operating condition l1
without and with the SVC stabilizer installed is presented in Fig. 6.20. At 0.5 s of
simulation, a three-phase short-circuit fault occurred at node 8 and was cleared in
0.1 s. From the simulation result, it can be seen that the low-frequency oscillation is
well damped.
To examine the robustness of the SVC stabilizer, the oscillation mode of interest
is calculated at other operating conditions l2 l10 with the SVC stabilizer
installed, and the results are given in Table 6.10. From Table 6.10, it can be
observed that when the SVC stabilizer is designed at the selected operating condition l1 which is predicted to be least effective, its effectiveness is ensured when
the system operates at other operating conditions. Hence, the design ensures the
robustness of the stabilizer.
320
6.4
Example Three-Machine Power System
6.4.1
Dynamic Interactions Among PSSs Installed

in Example Three-Machine Power System
In Sect. 5.3, an example three-machine power system is presented. The demonstration of the example is nished with the coordinated design of three PSSs
installed in three synchronous generators in Sect. 5.3.2.3, where the forward path of
the PSSs before they are designed and installed is calculated as

K2 KA K k
T0 I k
TA K6 KA 1
Fpss k
1
1 d0
1
3
3
2
4:0590\29:34 1:7719\153:70 1:0376\155:74
7
6
4 0:6433\113:76 6:1802\28:79 2:6963\136:09 5
1:1801\125:70 2:6184\140:40
5:3227\31:52
0:7094 j7:0590 is the target oscillation mode for the coordinated
where k
1
1
design of the PSSs, i.e. the position of the oscillation mode k
0:1736 j7:0590 to be assigned to by the coordinated design of the PSSs. With
the PSSs being designed by the use of the method introduced in Sect. 5.2.2, the
. The mode with the
oscillation mode is not moved to the exact target position, k
1
pss1 0:6094 j7:0235. The eigenvalue drift
PSSs installed is in fact moved to k
to k
pss1 0:6094 j7:0235 is due to the dynamic
from the target position k
1
interactions between the PSSs as to be explained as follows.
Consider the example three-machine power system with the PSSs installed on
each of the synchronous generators. Transfer function of the PSSs is
Tpssi s Kpssi
1 sT2i 1 sT4i
; T1i T3i 0:05; i 1; 2; 3
1 sT1i 1 sT3i
Parameters of the PSSs are obtained as given in Table 5.6 in Sect. 5.3.2.3. The
pss1 0:6094 j7:0235 can be obtained to be
forward path of the PSSs with k

pss1 K2 KA K k
pss1 T0 I k
pss1 TA K6 KA 1
Fpss k
d0
3
2
3
4:2427\29:72 1:8428\153:34 1:0765\155:48
6
7
4 0:7022\115:88 6:4424\29:37 2:8752\135:23 5
1:2780\125:43 2:7740\139:57 5:5693\32:24
According to Eq. (6.115), damping torque provided by each PSS to each of the
three generators can be calculated as
6.4 Example Three-Machine Power System
321
pss1 T
pss1 Dxi ReFpss k
pss1 Fi T
pss1 Dxi ; i
PSS k
pssi k
pssi k
DTPSSi ReF
1; 2; 3
where
2 3
2 3
2 3
1
0
0
F1 4 0 5; F1 4 1 5; F1 4 0 5
0
0
1
The following results can be obtained according to the above equations:
2
22:7134Dx1
3:5277Dx2
7
7
6
6
DTPSS1 4 3:0879Dx1 5; DTPSS2 4 12:3402Dx2 5;
6:1895Dx1
5:2042Dx2
3
2
4:4036Dx3
7
6
DTPSS3 4 11:5351Dx3 5
22:9518Dx3
The above results give the damping torque contribution from the ith PSS to the
jth generator in the form of Dpssij Dxi ; i; j 1; 2; 3; i 6 j. This needs to be converted to the form Dpssij Dxj ; i; j 1; 2; 3; i 6 j. The conversion is expressed
generally by Eq. (6.117). Without the PSS installed, the conversion coefcient
i can be calculated as explained from Eqs. (6.128) to (6.135) or from

cj k
Eqs. (6.150) to (6.154). With the PSSs installed, the conversion can be made as
explained as follows.
With the PSSs installed, the closed-loop state equation is
sDX Apss DX
which is given at the end of Sect. 5.3.2.3. Hence,
pss1 DX
pss1 Apss DX
pss1
k
k
k
pss1 to be
Denote the right eigenvector of the state matrix corresponding to k

vpss1 ; it should have
pss1 vi k
pss1
i k
DX

DXj kpss1 vj kpss1
pss1 and DX
pss1 is the ith and jth state variables (element) of
j k
i k
where DX
pss1 and vj k
pss1 are the ith and jth elements of vpss1 ,
pss1 , and
k
vi k
DX
respectively.
322
From the closed-loop state matrix obtained in Sect. 5.3.2.3, the right eigenvector
pss1 0:6094 j7:0235; vpss1 is calculated as
corresponding to k
vpss1 vpss11
vpss12

vpss13 T
where
2

vpss11
0:0159 j0:0289
0:0554 + j0:0831
3
0:0049 j0:0047
6 0:0042 j0:0198 7
7
6
7
6
6 0:0107 j0:0087 7
7
6
6 0:3141 j0:2922 7;
7
6
7
6
5
4
0:7337 j0
2
7
6
7
6
7
6
7
6
0:0163 j0:0429
7
6
6 6:1622 j4:1263 104 7; vpss12
7
6
7
6
4 1:7503 j1:3991 103 5
9:2666 j4:4722 104
3
0:0005 j0:0006
6 0:0016 j0:0035 7
7
6
7
6
6 0:0015 j0:0020 7
7
6
6
7
6 0:0015 j0:0039 7
7
6
4 0:0009 j0:0008 5
0:3582 j0:3680

vpss13
0:0021 j0:0035
DX is dened in Sect. 5.3.2.3 to be

DX DX1
DX2
DX 3
T
where
DX1 Dd1 Dd2 Dd3 Dx1 Dx2 Dx3 T

T
DX2 DE0q1 DE0q2 DE0q3 DE0fd1 DE0fd2 DE0fd3
DX3 DXpss1
DUpss1
DXpss2
DUpss2
DXpss3
DUpss3 T
Thus, from the last three elements of vpss11 , the following results can be obtained
as
4
pss1 Dx2 6:1622 j4:1263 10 Dx2 0:3298 j0:0279Dx2 ;
Dx1
c12 k
1:7503 j1:3991 103
4
pss1 Dx3 6:1622 j4:1263 10
c12 k
Dx3 0:7137 j0:1009Dx3 ;
Dx1
9:2666 j4:4722 104
3
pss1 Dx3 1:7503 j1:3991 10 Dx3 2:1230 j0:4853Dx3
c23 k
Dx2
9:2666 j4:4722 104
323
Hence, with Dpssij Dxi ; i; j 1; 2; 3; i 6 j being converted to Dpssij Dxj ; i; j

1; 2; 3; i 6 j, the damping torque contributed by each PSS to each of the generators
should be calculated as
2
DTPSSi
3
pss1 T
pss1 c k

pssi k
ReFpssi1 k
i1 pss1 Dx1
6
pss1 T
pss1 c k
pss1 Dx2 7
pssi k
4 ReFpssi2 k
5; i 1; 2; 3
i2
pss1 T
pss1 c k

pssi k
ReFpssi1 k
i3 pss1 Dx3
By using the results obtained and equation given above, the damping torque
contributed by each PSS to each of the generators is calculated as
2
22:7134Dx1
11:0506Dx1
7
7
6
6
DTPSS1 4 0:8905Dx2 5; DTPSS2 4 12:3402Dx2 5;
5:9790Dx3
14:7900Dx3
3
2
4:9442Dx1
7
6
DTPSS1 4 3:4809Dx2 5
22:9518Dx3
The above results are summarized in Table 6.11.
Sect. 5.3.2.3, the required damping torque provided by the PSSs is
In
Dpss1 Dpss2 Dpss3 p52 21:73 11:84 21:94 to move the oscillation
0:7094 j7:0590 by the design of
1 0:1736 j7:0590 to k
mode from k
1
the PSSs. With the PSSs installed, the oscillation mode is only moved to
pss1 0:6094 j7:0235. The reason is clearly indicated by the results in
k
Table 6.11. The actual damping torque provided by the PSSs is
Dpss1 Dpss2 Dpss3 38:71 9:75 14:14 , not the required damping torque

Dpss1 Dpss2 Dpss3 21:73 11:84 21:94 for the PSSs design, because
each PSS provides damping torque to all three generators as is shown in Table 6.11.
1. The amount of damping torque provided by each PSS to the generator where it is
installed
is
22:7134
12:3402 22:9518 ,
very
close
to

Dpss1 Dpss2 Dpss3 21:73 11:84 21:94 , the required damping torque
by the design of the PSS. The difference is caused by the difference between

k
pss1 0:6094 j7:0235 and k1 0:7094 j7:0590.
Table 6.11 The damping torque provided by each PSS

Damping torque contribution to
By PSS1
By PSS2
By PSS3
Total Dpssi
G1
G2
G3
22.7134
0.8905
5.9790
11.0506
12.3402
14.7900
4.9442
3.4809
22.9518
38.7082
9.7498
14.1407
324
2. Each of the three PSSs provides positive damping torque to the rst generator.
Thus, the rst generator obtains more damping torque (38:7082Dx1 ) than
required by the design.
3. PSS installed on the second generator and third generator contributes negative
damping torque to the third and second generators, respectively, reducing the
actual amount of damping torque obtained by the second and third generators.
6.4.2
Design of Non-negatively Interactive PSSs Installed

in the Example Power System
In Sect. 6.2.4.2, a scheme of non-negatively interactive design of the PSSs is

introduced. In this section, the scheme is used to examine whether the coordinated
design of three PSSs to be installed in the three-machine power system can be
designed to be non-negatively interactive or not as follows.
The state matrix of the example three-machine power system is constructed in
Sect. 5.3.1.3 by the following coefcient matrices according to Eq. (5.43) when the
PSSs are not installed
2
2:3373
6
K1 4 1:5290
1:0599
2
1:1889
6
K3 4 0:1209
1:4640
2:2160
0:8201
0:4521
1:2478
0:1067
3:1565
3
3
2
0:8733
2:4574 0:9034 0:5628
7
7
6
0:6870 5; K2 4 1:0247 3:0671 0:4060 5;
1:8800
0:3694 0:6458 2:3234
3
3
2
0:0994
0:0277 0:0008 0:0269
7
7
6
0:7660 5; K4 4 1:2024 1:8725 0:6701 5;
3:7141
1:2795 1:0112 2:2907
3
3
2
0:0068 0:0076 0:0143
0:8732 0:0711 0:0678
7
7
6
6
K5 4 0:0596 0:0315 0:0281 5; K6 4 0:4315 0:5292 0:1730 5;
00780 0:0499 0:1278
0:4691 0:1829 0:4790
3
2
3
2
0 0 0
47:2
0
0
7
6
7
6
M4 0
12:8
0 5; D 4 0 0 0 5;
2
0 0 0
0
0
6:02
3
3
2
200
0
0
0:02
0
0
7
7
6
6
KA 4 0
200
0 5; T A 4 0
0:02
0 5;
0
0
200
0
0
0:02
3
2
8:96
0
0
7
6
T0d0 4 0
6:00
0 5
0
0
5:89
2
In Sect. 5.3.2.3, it is found that when the damping torque provided

by the PSSs

installed on the three generators is Dpss1 Dpss2 Dpss3 p52 21:73
11:8421:94, the electromechanical oscillation mode of interests is moved to
0:7094 j7:0590. Hence, the state matrix of the example three-machine
k
1
325
power system with the PSSs installed (parameters of the PSSs have not been set yet)
can be constructed still according to Eq. (5.43) with coefcient matrix D being
changed to
2
Dpss1
4
0
D
0
Dpss2
0
3 2
0
21:73
0 54 0
Dpss3
0
0
11:84
0
3
0
0 5
11:84
Denote the state equation with the PSSs installed (their parameters have not been
set yet) to be
~ pss DX
~ A
~
sDX
~ DX1
where DX
DX2 T
DX1 Dd1 Dd2 Dd3 Dx1 Dx2 Dx3 T

T
DX2 DE0q1 DE0q2 DE0q3 DE0fd1 DE0fd2 DE0fd3
0:7094 j7:0590 is the eigenvalue of state matrix A
~ pss . Let
Obviously, k
1

and v0 k

~ pss corresponding to k

v0pss1 be the right eigenvector of state matrix A
j 1 be
1
the (j + 3) element of v0pss1 ; j 1; 2; 3. It can have

Dxj vi k1 Dxj
Dxi cij k
1

vj k
1
The damping torque provided by the ith PSS to the jth generator is
c k

DTpssij ReFpssij k
1 ij 1 Tpssi k1 Dxj ; i; j 1; 2; 3

f pssij Fpssij k
1 cij k1 ; i; j 1; 2; 3 is the forward path from the ith PSS to the

k
electromechanical oscillation loop of the jth generator. A forward path matrix F
1
c k
; i; j 1; 2; 3 being the elements. For the
can be formed with f pssij Fpssij k
1 ij 1
example three-machine power system, v0pss1 is calculated as

v0pss1
0pss1a
v
0pss1b
v
v0pss1c
0

v0pss1b
v0pss1c T
vpss1a

0:0080 j0:0556 0:0111 + j0:1722 0:0188 j0:0752 0:0013 j0:0001 T
0:0038 j0:0006 0:0017 + j0:0003 0:0014 + j 0:0014 0:0086 j0:0222 T
0:0076 j0:0021 0:0958 + j0:0822 0:9324 + j0 0:16467 j0:2189 T
326
is calculated as
Thus,
cij k
1
0
c12
v0 pss14
v pss15 0:0013 j0:0001=0:0038 j0:0006 0:3187 + j0:0671;
0
0
pss14
v pss16 0:0013 j0:0001=0:0017 + j0:0003 0:7211 j0:0733;
c13 v
c21 1=
c12 1=0:3187 + j0:0671 3:0048 j0:6331;
0
c23 v
0 pss15
v pss16 0:0038 j0:0006=0:0017 + j0:0003 2:1204 + j0:6768;
c13 1=0:7211 j0:0733 1:3726 + j0:1395;
c31 1=
c23 1=2:1204 + j0:6768 0:4280 j0:1366
c32 1=
is calculated as
k
F
1
2

k
F
1
4:059\29:341
4 0:2095\54:347
0:8554\60:103
5:441\38:198
6:1802\28:783
5:8279\122:7
3
1:4316\18:46
1:2114\153:79 5
5:3227\31:527
c k

Three vectors f pssij Fpssij k
1 ij 1 ; j 1; 2; 3 for i = 1,2,3 are plotted in
Figs. 6.21, 6.22, and 6.23, respectively.
From Fig. 6.21, it can be seen that if the compensating phase of PSS1 is chosen
Tpss1 \h10 Tpss1 \29:341 . The
pss1 k
to be h10 29:341 , it can have T
1
electric torque provided by PSS1 to each of the three generators is
Electric torque
provided to G3
forward
path to G2
-1
0
3
Electric torque
provided to G1
10
Electric torque
provided to G2
forward path
to G3
-1
Direction of positive
synchronizing torque
-2
Fig. 6.21 Forward path of PSS1 to each generator
forward path to G1
327

c k

DTPSS11 Fpssi11 k
1 i1 1 Tpss1 k1 Dx1 4:059\29:341 Tpss1 \h10 Dx1 4:059Tpss1 Dx1

c k

DTPSS12 Fpssi12 k
1 i1 1 Tpss1 k1 Dx2 0:2095\54:347 Tpss1 \h10 Dx2
0:2095Tpss1 cos54:347 29:341 Dx2
c k
T
Dx3 0:8554\60:103 Tpss1 \h10 Dx3
pss1 k
DTPSS13 Fpssi13 k
1
i1
0:8554Tpss1 cos60:103 29:341 Dx3
which are all positive. The synchronizing torque provided by PSS1 to each of the
generators is

c k

DTs11 jImfFpssi11 k
1 i1 1 Tpss1 k1 gDx1
DTs12
DTs13
j4:059Tpss1 sin29:341 h10 Dx1 0

j0:2095Tpss1 sin54:347 29:341 Dx1
jx
j0:2095Tpss1 sin54:347 29:341 s Dd2
x0
x
0:2095Tpss1 sin54:347 29:341 s Dd2
x0
x
0:8554Tpss1 sin60:103 29:341 s Dd3
x0
It shows that PSS1 provides positive synchronizing torque to the 2nd and 3rd
generators, as the projection of the vector of electric torque provided by PSS1 to G2
and G3 is on the direction of positive synchronizing torque.
From Fig. 6.22, it can be seen that if the compensating phase of PSS2 is chosen
to be h20 , PSS2 will provide positive damping and synchronizing torque to G1
and G2, but negative damping torque to G3. There is no suitable compensating
forward path to G3
4
Electric torque
provided to G3
2
Electric torque
-2
2
0
4 provided to G2 6
20
Electric torque
provided to G1
-2
forward path to G2
forward path to G1
-4 Direction of positive
328
forward path to G2
-1
Electric torque
provided to G1
1
0
Electric torque
provided to G2
Electric torque
provided to G2
30
forward path to G1
-1
-2
Direction of positive
forward path to G3
phase to be selected to ensure the PSS to provide positive damping torque to three
generators. Similarly, from Fig. 6.23, it can also be seen that there is no suitable
compensating phase available for designing PSS3. Hence, in the example
three-machine power system presented in Sect. 5.3, it is inevitable that PSS2 and
PSS3 supply negative to other generators.
References
1. Wang HF, Swift FJ (1998) A unied model for the analysis of FACTS devices in damping
power system oscillations part II: multi-machine power systems. IEEE Trans Power Deliv (4)
2. Wang HF, Swift FJ (1997) The indexes for selecting the best locations of PSS or FACTS-based
stabilizers in multi-machine power systems: a comparison study. IEE Proc Part C (2)
3. Wang HF, Swift FJ (1997) The connection between modal analysis and electric torque analysis
in studying the oscillation stability of multi-machine power systems. Int J Electr Power Energy
Syst 19(5):321330
4. Wang HF (1999) Selection of robust installing locations and feedback signals of FACTS-based
stabilizers in multi-machine power systems, IEEE Trans Power Syst (2)
5. Wang HF (1999) Design of non-negatively interactive FACTS-based stabilizers in
multi-machine power systems. Int J Power Syst Res 50 (3)
Chapter 7
Multi-machine Power Systems Installed

with VSC-Based Stabilizers
7.1

System Installed with VSC-Based Stabilizers
7.1.1

System Installed with a Shunt VSC-Based Stabilizer
7.1.1.1

Installed with a Shunt VSC-Based Stabilizer
Without loss of generality, it can be assumed that a shunt VSC is installed at a busbar
between nodes 1 and 2 in an N-machine power system, as shown in Fig. 7.1. The
shunt VSC can be connected to an energy storage system (ESS), a HVDC, or a static
renewable power station, such as a photovoltaic (PV) or fuel cell (FC) power station.
It can operate alone as a VSC-based FACTS device, STATCOM (see, Sect. 4.1.1.1).
From Fig. 7.1, it can be obtained by the following:
s jxsIs V
c jxs I1s Is2 V
c
V
7:1
Hence,
1 jx1sI1s V
s jx1sI1s jxs I1s Is2 V
c jx1s xs I1s jxsIs2 V
c
V
2 jxs2Is2 V
s jxs2Is2 jxs I1s Is2 V
c jxsI1s jxs2 xs Is2 V
c
V
7:2
In matrix form, Eq. (7.2) is given as follows:

I1s

jx1s xs
jxs
V
V
1 c

jxs
jxs2 xs Is2
V2
Vc

DOI 10.1007/978-1-4899-7696-3_7
7:3
329
330
Multi-machine Power Systems Installed with VSC-Based Stabilizers
Fig. 7.1 A multi-machine

shunt VSC
A multi-machine power system
V1
x 1s
Vs
x s2
V2
Is2
I1s
Node 1
Node 2
Is
xs
VC
Shunt VSC
That is,
I1s
Is2
jx1s xs
jxs
jxs
jxs2 xs
1

1

V
V
c

V2
Vc
7:4
For the N-machine power system, rstly the set of network equations with N
generators internal nodes, nodes 1 and 2 left (without node s where the shunt VSC
is installed) can be established as follows:
3 2
0
y11
6 7 6
4 0 5 4 y21
Ig
31
Y
2
y12
y22
32
Y
32 3
1
13
V
Y
7
6
2 7
23 54 V
Y
5

Eg
Y33
7:5
Secondly, the network equations of (7.5) can be modied with node s added to be
0

y11 0
0
V1
I1s
Y13
E
23 g
2
Is2
0 y022 V
0
Y
7:6

V1
Ig Y

31 Y
32
Y33 Eg
2
V
y022 are obtained by excluding x12 x1s xs2 from y11 and y22 in
where
y011 and
Eq. (7.5), respectively.
Finally from Eqs. (7.4) and (7.6), it can have the following:

0
0

y011 0
V1
0
2
0 y22 V

1
jx1s xs
jxs
V1
Vc
Y13
E
23 g

jxs
jxs2 xs
V2
Vc
Y
7.1 Mathematical Model of a Multi-machine Power System
331
That is.

1
0
y11 0
jxs
jx1s xs
V1
I
2
0 y022
jxs
jxs2 xs
V

jx1s xs
jxs
Vc
Y13

E
23 g
c
jxs
jxs2 xs Y
V

jx1s xs
jxs
Y13
1 Vc
E
Y
S
23 g

jxs
jxs2 xs Y
Vc
7:7
Substituting Eq. (7.7) into (7.6), it can be obtained that

1 1
V

Ig C
E g Y
c
c Y
31 Y
32 Ys
V
1

1 jx1s xs

33 Y
31 Y
32 Y
Y
s
jxs
jxs
jxs2 xs

Y13
g
E
23
Y
7:8
From Eqs. (7.4) and (7.7), it can have the following:

I1s
Is2
jx1s xs
jxs
jxs
jxs2 xs
1

13
Vc
1
1 Y

E
YS I YS
7:9
Vc
Y23 g
Hence from Fig. 7.1 and Eq. (7.9), it can have the following:
Is isx jisy I1s Is2 1

N
X
I
c
gj
1 1s ys V
ysj E
Is2
j1
7:10
g . According to Eq. (5.10), E

gj is the jth element of E
gj is given as follows:
where E
gj E0 ejdj xqj x0 iqj eidj 90
E
qj
dj
7:11

s jxsIs V
c jxs ys V
c
V
N
X
gj
ysj E
7:12
j1
In xy coordinate of the power system, it can have the following:

c mkVdc cos c j sin c mkVdc \c
V
7:13
c and x-axis. Active power received by the capacitor

where c is the angle between V
of the VSC from the power system is given as:
332
Vdc Idc1 isx vcx isy vcy isx mkVdc cos c isy mkVdc sin c
Hence,
Idc1 isx mk cos c isy mk sin c
7:14
In this section, for the simplicity of discussion, effect of dynamic and control
functions associated with Idc2 (see, Fig. 4.1) is not considered, and the VSC adopts
the PWM (pulse width modulation) algorithm. Hence, the dynamic equation on the
DC side of the VSC is given as follows [1]:
1
1
V_ dc
Idc1
isx mk cos c isy mk sin c
Cdc
Cdc
7:15
AC and DC voltage control functions implemented via the DC/AC converter are
as follows:
m m0 Tm sVs Vsref uvscpss
/ /0 Tdc sVdc Vdcref
7:16
s and V
c
where uvscpss is the stabilizing control signal, and / is the angle between V
ac tan1
vsy
/c
vsx
7:17
Tm s) and Tdc s) are the transfer functions of AC and DC voltage control

function of the VSC, respectively.
and Y
in Eq. (7.8) as
Denote the elements of matrix C
dci e;
C
l m
yij
Y
7:18
Similar to Eqs. (6.6) and (6.7), from Eqs. (7.8) and (7.18) it can have the
following:
Igi idi jiqi ixi jiyi ej90 di
!
N

h
i
X

j
d
90
j

0
jd
0
c
yij Eqj e j xqj xdj iqj e j
e 90 di ci V
j1
ej90
di
c
ci V
N
X
h

i

yij E0qj ej90 dj di xqj x0dj iqj ejdj di
j1
ci mkVdc ej90
di c bi
N
X
h

i

yij E0qj ej90 aij dj di xqj x0dj iqj ejaij dj di
j1
7:19
333
where ci ci ejbi ;

yij yij ejaij . From Eq. (7.19), it can be obtained that
idi ci mkVdc sinc bi di
N
h
i

X
yij E0qj sinaij dij xqj x0dj iqj cosaij dij

j1
iqi ci mkVdc cosc bi di

N
h
i

X
7:20
j1
Linearization of Eq. (7.20) can be written in the matrix form to be

DId Fdd Dd Gdd DE0q Hdd DIq Ldddc DVdc Lddm Dm Lddg Dc
DIq Fqq Dd Gqq DE0q Hqq DIq Lqqdc DVdc Lqqm Dm Lqqg Dc
7:21
That can give the following:
DId Fd Dd Gd DE0q Lddc DVdc Ldm Dm Ldg Dc
DIq Fq Dd Gq DE0q Lqdc DVdc Lqm Dm Lqg Dc
7:22
Substituting Eq. (7.22) into Eqs. (5.40) and (5.41), it can have the following:
Dd_ xo Dx
Dx_ M1 K1 Dd K2 DE0q DDx Kpdc DVdc Kpm Dm Kpg Dc
0
0
0
DE_ q T01
d0 K3 DEq K4 Dd DEfd Kqdc DVdc Kqm Dm Kqg Dc
0
0
1
0
DE_ fd T1
A DEfd TA KA K5 Dd K6 DEq Kvdc DVdc Kvm Dm Kvg Dc
7:23
Linearization of Eqs. (7.10) and (7.11) is given as follows:
Disx axdc DVdc axm Dm axg Dc axd Dd axe DE0q
Disy aydc DVdc aym Dm ayg Dc ayd Dd aye DE0q
7:24

DV_ dc K7 Dd K8 DE0q K9 DVdc K10 Dm K11 Dc
7:25
334
Linearization of Eqs. (7.11) and (7.12) can give the following:

Dvsx bxdc DVdc bxm Dm bxg Dc bxd Dd bxe DE0q
Dvsy bydc DVdc bym Dm byg Dc byd Dd bye DE0q
Hence, linearization of Eq. (7.16) is given as
Dm Tm sDVs Duvscpss
Tm sbdc DVdc bm Dm bg Dc bd Dd be DE0q Duvscpss
7:26
D/ Tdc sDVdc
That is,
Dm Tm sDVs Duvscpss
Tm s
bdc DVdc bg Dc bd Dd be DE0q Duvscpss
1 bm Tm s
7:27
D/ Tdc sDVdc
Dc Kgd Dd Kgq DE0q Kgdc DVdc Kgm Dm Kgf D/
7:28
By substituting Eq. (7.28) into Eqs. (7.23), (7.25), and (7.27), HeffronPhillips
model of the power system installed with the shunt VSC-based stabilizer can be
obtained as
Dd_ xo Dx
Dx_ M1 K01 Dd K02 DE0q DDx K0pdc DVdc K0pm Dm K0pg D/
0
0
0
0
0
0
0
0
DE_ q T01
d0 K3 DEq K4 Dd DEfd Kqdc DVdc Kqm Dm Kqg D/
0
0
1
0
0
0
0
0
0
DE_ fd T1
A DEfd TA KA K5 Dd K6 DEq Kvdc DVdc Kvm Dm Kvg D/
DV_ dc K07 Dd K08 DE0q K09 DVdc K010 Dm K011 D/
Dm
Tm s
b0 DVdc b0g D/ b0d Dd b0e DE0q Duvscpss
1 bm Tm s dc
D/ Tdc sDVdc
7:29
335
The model is shown in Figs. 7.2 and 7.3 [2, 3].
bdc
+
be
Eq '
Tm (s)
1 b m Tm (s)
bd
K 8
1
s K9
K10
Vdc
u vsc pss
K 7
Tdc (s)
K11
bg
Fig. 7.2 HeffronPhillips modelpart of shunt VSC
K 1
0 I
s
(sM + D) 1
K 4
K 5
Vdc
[K p-dc K p-g K p-m ]
K 2
[]
m
[K v-dc K v-g K v-m ]
[K q-dc K q-g K q-m ]

+
Eq '
(sTd0' + K 3 )
Efd'
K 6
Fig. 7.3 HeffronPhillips modelpart of power system
(I + sTA ) K A
336
7.1.1.2
General Linearized Model of an N-Machine Power System

Installed with a Shunt VSC-Based Stabilizer
Let DVsxy

Dvsx
. Linearization of Eq. (7.17) can be obtained as

Dvsy
Dc D/ aTs DVsxy
7:30

vsx jvsy jxs isx jisy mkVdc cos c jmkVdc cos c
7:31
By using Eq. (7.30), linearization of Eq. (7.31) can give the following:
Disx cxdc DVdc cxm Dm cxf D/ aTsx DVsxy
Disy cydc DVdc cym Dm cyf D/ aTsy DVsxy
7:32
Hence, by using Eq. (7.32), linearization of Eq. (7.15) can be obtained as

DV_ dc cddc DVdc cdm Dm cdf D/ aTsdc DVsxy
7:33
Because
h
DVs vVsx0
s0
i
vsy0
Vs0 DVsxy ;
7:34
the state-space realization of the AC controller of the shunt VSC (the rst equation
in Eq. (7.26)) can be written as
X_ AC AAC XAC BAC DVsxy
Dm CAC XAC DAC DVsxy Duvscpss
7:35
Let the state-space realization of the DC controller of the shunt VSC (the second
equation in Eq. (7.26)) be
X_ DC ADC XDC BDC DVdc
D/ CDC XDC DDC DVdc
7:36
From Eqs. (7.33), (7.35), and (7.36), it can have the following:
DV_ dc cddc cdf DDC DVdc cdm CAC XAC cdf CDC XDC
cdm DAC aTsdc DVsxy cdm Duvscpss
7:37
337
From Eqs. (7.35)(7.37), the overall state equation of the shunt VSC dynamics
and control functions can be written as
X_ VSC AVSC XAC BVSC DVsxy Bvscpss Duvscpss
7:38
where
3
DVdc
7
6
4 XAC 5;
2
XAC
AVSC
cddc cdf DDC

6
4
0
BDC
XDC
2
BVSC
6
4
cdm CAC
AAC
cdm DAC aTsdc

BAC
0
3
7
5;
0
2
Bvscpss
3
cdf CDC
7
0
5;
ADC
cdm
7
6
4 0 5
0
By using Eqs. (7.35) and (7.36), Eq. (7.32) becomes as

Disx cxdc cxf DDC DVdc cxm CAC XAC cxf CDC XDC
cxm DAC aTsx DVsxy cxm Duvscpss
Disy cydc cyf DDC DVdc cym CAC XAC cyf CDC XDC
7:39
cym DAC aTsx DVsxy cym Duvscpss

Let DIsxy

Disx
. From Eq. (7.39), it can have the following:

Disy
DIsxy CVSC XVSC DVSC DVsxy Dvscpss Duvscpss
7:40
where
CVSC
DVSC

cxdc cxf DDC cxm CAC cxf CDC
;
cydc cyf DDC cym CAC cyf CDC
"
#

cxm DAC aTsx
cxm
; Dvscpss
cym DAC aTsy
cym
If it is assumed that the feedback signal of the shunt VSC-based stabilizer is the
active power, Ps2 , delivered along the transmission line where the shunt VSC is
installed, similar to Eq. (6.40), it can have the following:
DPs2 psvsc DVsxy p2vsc DVxy2
7:41
338

Dvx2
2 . Similar to Eq. (6.98), let the linwhere DVxy2
and vx2 jvy2 V
Dvy2
earized network equation of the N-machine power system with the shunt
VSC-based stabilizer installed be (only with N nodes of generators, node s and 2 of
system of Fig. 7.1 left) the following:
3 2
Ygg
DIxy
4 DIsxy 5 4 Ysg
Y2g
0
2
Ygs
Yss
Y2s
32
3
DVxy
Yg2
Ys2 54 DVsxy 5
Y22
DVxy2
7:42
By substituting Eq. (6.29) and (7.40) into Eq. (7.42), it can have the following:
3 2
Ygg Dgxy
Cgxy Xg
4 CVSC XVSC Dvscpss Duvscpss 5 4
Ysg
Y2g
0
2
Ygs
Yss DVSC
Y2s
32
3
DVxy
Yg2
Ys2 54 DVsxy 5
Y22
DVxy2
7:43

DV2xy

DVxy
Y2s
DVsxy

0
Cgxy
0
Xg
Duvscpss
Dvscpss
0
CVSC XVSC

DVxy
Ygs
Ygg Dgxy
Yg2 1

Y22 Y2g Y2s
Ysg
Yss DVSC
DVsxy
Ys2

DVxy
Ygsxy
DVsxy
Y1
22 Y2g
7:44
By arranging the state equation of the generators of Eq. (6.29) and that of the
shunt VSC of Eq. (7.38) together, it can have the following:
X_ g
X_ VSC
Agxy
0
0
AVSC

Bgxy
Xg
0
XVSC
BVSC

0
DVxy
Duvscpss
Bvscpss
DVsxy
7:45
By substituting Eq. (7.44) into (7.45), the open-loop state equation of the system
without the shunt VSC-based stabilizer installed can be obtained as
X_ AX BDuvscpss
7:46
339
where

0
0
Xg
Bgxy
0
1
X
; B
Y
Bvscpss
0
BVSC gsxy Dvscpss
XVSC

Agxy
0
0
Cgxy
0
Bgxy
1
A
Y
0
AVSC
0
BVSC gsxy 0
CVSC
DPs2 0
psvsc
psvsc p2vsc Y1

22 Y2g
0

p2vsc Y1
22 Y2g
Cgxy
0
0
CVSC
Y2s

DVxy
DVsxy

Y2s Y1
gsxy

0
Xg
Duvscpss
Dvscpss
XVSC
7:47
Hence, the output equation of the system is given as

Dy CX DDuvscpss
where
C 0
D 0
psvsc p2vsc Y1

22 Y2g
psvsc p2vsc Y1
22 Y2g
Y2s Y1
gsxy
Y2s Y1
gsxy
7:48
Cgxy
CVSC
Dvscpss
7.1.2

System Installed with a UPFC-Based Stabilizer
7.1.2.1

Installed with a UPFC-Based Stabilizer [4, 5]
Without loss of generality, it can be assumed that a UPFC is installed at a busbar

between nodes 1 and 2 in an N-machine power system, as shown in Fig. 7.4.
From Fig. 7.4, it can be obtained (see, Sect. 4.3.1.1) as
Et jxEIE V
E;
V
Bt jxBIE2 V
B
V
7:49
340
Ign
Gn
I g2
Vgn
G2
I g1
Vg2
G1
Vg1
Yt
V1
VEt
I1E
VBt
I E2
V2
xB
x1E
x E2
IE
VEt x E
Cdc
mE E
mB B
UPFC
Fig. 7.4 A multi-machine power system installed with a UPFC
and
Et
1 jx1EI1E V
V
Et jxE2IE2 V
Bt V
2
V
IE I1E IE2
7:50
Substituting Eq. (7.49) into (7.50), it can have the following:

I1E
IE2

1 jxE xE2 xB
jxE
xR
jxE
jx1E xE

1
1 jxE2 xB
V
2
jxE
V
xR
jxE
jx1E xE
E
V
B
V
7:51
where xR x1E xE xE xE2 xB x2E . Similar to Eq. (7.6), it should have the
following:

0
y11 0
0
V1
I1E
Y13
E
0
23 g

0 y22 V2
0
IE2
Y

V
1
Ig Y
g
33 E
31 Y
32
Y
2
V
7:52
341
022 are obtained by excluding x12 x1E xE2 from y11 and y22 in
where
y011 and y
Eq. (7.5). Substituting Eq. (7.51) into (7.52), it can be obtained that

y011 0
jxE
V1
1 jxE xE2 xB
0
2
x
0 y22
jxE
jx1E xE
0
V
R

jxE
VE
Y13
1 jxE2 xB
E
23 g
B
xR
jxE
jx1E xE V
Y

0
7:53
The above equation gives the following:

1
V

V2
1
Y
E

1 jxE2 xB
jxE
xR
jxE
jx1E xE

13
E
V
Y
23 Eg
B
V
Y
7:54

0

11 0
jxE xE2 xB
jxE
1
E y
where Y
xR
0 y022
jxE
jx1E xE
EV
BV
Ig C
E g
E C
B Y

1 1 jxE2 xB
jxE
VE

31 Y
32 Y
Y
E
B
xR
jxE
jx1E xE V

1 Y13
g
33 Y

31 Y
32 Y
E
Y
E
23
Y
7:55

jxE
I1E
1 jxE xE2 xB
1

Y I

IE2
xR
jxE
jx1E xE E

jxE
VE
1 jxE2 xB

B
xR
jxE
jx1E xE V

jxE
1 jxE xE2 xB
1 Y13

E

Y
23 g
xR
jxE
jx1E xE E Y
Hence from Fig. 7.4 and Eq. (7.56), it can have the following:

N
X
IE iEx jiEy I1E IE2 1 1 I1E yE V
E
B
gj

yB V
yEj E
IE2
j1

N
X
IE1 iE1x jiE1y 1 0 I1E yE1 V
E yB1 V
B
gj

yE1j E
IE2
j1

N
X
IE2 iE2x jiE2y 0 1 I1E yE2 V
E yB2 V
B
gj

yE2j E
IE2
j1
7:56
7:57
342
g , which is given by Eq. (7.11). From Eqs. (7.49)

gj is the jth element of E
where E
and (7.57), it can have the following:
Et jxE yE 1V
B jxE
E jxE yB V
V
N
X
gj
yEj E
j1
Bt jxB yE2 V
E jxB yB2 1V
B jxB
V
N
X
7:58
gj
yE2j E
j1
In xy coordinate of the power system, let the following be denoted

E mE kE Vdc cos cE j sin cE mE kE Vdc \cE
V

VB mB kB Vdc cos cB j sin cB mB kB Vdc \cB
7:59
Linearization of Eqs. (7.11) and (7.12) can give the following:

DvEtx bEtxdc DVdc bEtxmE DmE bEtxgE DcE
bEtxmB DmB bEtxgB DcB bEtxd Dd bEtxe DE0q
DvEty bEtydc DVdc bEtymE DmE bEtygE DcE
bEtymB DmB bEtygB DcB bEtyd Dd bEtye DE0q
DvBtx bBtxdc DVdc bBtxmE DmE bBtxgE DcE
bBtxmB DmB bBtxgB DcB bBtxd Dd bBtxe DE0q
7:60
DvBty bBtydc DVdc bBtymE DmE bBtygE DcE

bBtymB DmB bBtygB DcB bBtyd Dd bBtye DE0q
E to
Et and V
In xy coordinate of the power system, denote the angle between V
Bt and V
B to be dB ; that is,
be dE and that between V
vEty
dE cE
vEtx
vBty
ac tan1
dB cB
vBtx
ac tan1
7:61
From Eq. (7.60) and linearization of Eq. (7.61), it can be obtained that
DcE KgEd Dd KgEq DE0q KgEdc DVdc
KgEmE DmE KgEdE DdE KgEmB DmB KgEdB DdB
DcB KgBd Dd KgBq DE0q KgBdc DVdc
KgBmE DmE KgBdE DdE KgBmB DmB KgBdB DdB
7:62
343
E, C
B , and Y
in Eq. (7.55) as
Denote the elements of matrix, C
E dcEi e;
C
l m
yij
Y
B dcBi e;
C
7:63

Igi idi jiqi ixi jiyi ej90 di
ej90
di
E cBi V
B
cEi V
N
X
j1
ej90
di
E cBi V
B
cEi V
N
X
yij E0qj ej90
dj di
j1
cEi mE kE Vdc ej90
N
X
yij E0qj ej90
di cE bEi
cBi mB kB Vdc ej90
aij dj di
di cB bBi
j1
7:64
where cEi cEi ejbEi ; cBi cBi ejbBi ; yij yij ejaij . From Eq. (7.64) it can be obtained
that
idi cEi mE kE Vdc sincE bEi di cBi mB kB Vdc sincB bBi di
N
X
yij E0qj sinaij dij xqj x0dj iqj cosaij dij
j1
iqi cEi mE kE Vdc coscE bEi di cBi mB kB Vdc coscB bBi di
N
X
7:65
j1
Linearization of Eq. (7.65) is given as

DId Fdd Dd Gdd DE0q Hdd DIq Ldddc DVdc
LddmE DmE LddgE DcE LddmB DmB LddgB DcB
DIq Fqq Dd Gqq DE0q Hqq DIq Lqqdc DVdc
LqqmE DmE LqqgE DcE LqqmB DmB LqqgB DcB
7:66
344
By using Eq. (7.62), from above equation it can have the following:
DId Fd Dd Gd DE0q Lddc DVdc
LdmE DmE LddE DdE LdmB DmB LddB DdB
DIq Fq Dd Gq DE0q Lqdc DVdc
7:67
LqmE DmE LqdE DdE LqmB DmB LqdB DdB

Substituting Eq. (7.67) into Eqs. (5.40) and (5.41), it can be obtained that
Dd_ xo Dx
Dx_ M1 K1 Dd K2 DE0q DDx Kpdc DVdc
KpmE DmE KpdE DdE KpmB DmB KpdB DdB
0
DE_ q
0
0
T01
d0 K3 DEq K4 Dd DEfd Kqdc DVdc
7:68
KqmE DmE KqdE DdE KqmB DmB KqdB DdB

0
DE_ fd
0
1
0
T1
A DEfd TA KA K5 Dd K6 DEq Kvdc DVdc
KvmE DmE KvdE DdE KvmB DmB KvdB DdB

Dynamic equation of the VSCs of the UPFC on the DC side is given by the
following equation (see Eq. (4.73)):

_Vdc 1 mE kE VEt sin dE mB kB VBt sin dB
Cdc
xE
xB
7:69
By using Eqs. (7.60) and (7.62), linearization of Eq. (7.69) can be obtained as
DV_ dc K7 Dd K8 DE0q K9 DVdc
K10E DmE K11E DdE K10B DmB K11B DdB
7:70
Without loss of generality and for the purpose of demonstration, it is assumed

that the shunt VSC of the UPFC implements the AC and DC voltage control
functions, the series VSC realizes the line active and reactive power regulation, and
the damping control is added on the AC voltage control function as follows:
mE mE0 Tac sVEt VEtref uUPFCpss
dE dE0 Tdc sVdc Vdcref
mB mB0 Tap sPEt PEtref
dB dB0 Trp sQEt QEtref
7:71
345
where uUPFCpss is the damping control signal, Tac s), Tdc s),Tap s), and Trp s) is
the transfer function of the AC voltage, DC voltage, active, and reactive power
controller of the UPFC, respectively. Linearization of Eq. (7.71) is given as
DmE Tac sDVEt DuUPFCpss
DdE Tdc sDVdc
7:72
DmB Tap sDPEt

DdB Trp sDQEt
By using Eqs. (7.60) and (7.62), it can be obtained that
DVEt bEtdc DVdc bEtmE DmE bEtdE DdE
bEtmB DmB bEtdB DdB bEtd Dd bEte DE0q
7:73

"
#
N
X

yEj Egj
PEt jQEt VEt I jxE yE 1VE jxE yB VB jxE
E1
j1
"

yE1 V
E

yB1 V
B
N
X
7:74

yE1j E
gj
j1
By using Eq. (7.62), linearization of the above equation can be obtained as

DPEt bpdc DVdc bpmE DmE bpdE DdE
bpmB DmB bpdB DdB bpd Dd bpe DE0q
DQEt bqdc DVdc bqmE DmE bqdE DdE
bqmB DmB bqdB DdB bqd Dd bqe DE0q
7:75
Substituting Eqs. (7.73) and (7.75) into Eq. (7.72) gives the following:
DmE
Tac s
bEtdc DVdc bEtdE DdE bEtmB DmB
1 bEtmE Tac s
1
Dupss
bEtdB DdB bEtd Dd bEte DE0q
1 bEtmE Tac s
DdE Tdc sDVdc

Tap s
bpdc DVdc bpmE DmE
DmB
1 bpmB Tap s
bpdE DdE bpdB DdB bpd Dd bpe DE0q
DdB
Trp s
bqdc DVdc bqmE DmE
1 bqdB Trp s
bqdE DdE bqmB DmB bqd Dd bqe DE0q
7:76
346
Equation (7.68), (7.70), and (7.76) are the HeffronPhillips model of the power
system installed with the UPFC.
7.1.2.2
General Linearized Model of an N-Machine Power System

Installed with a UPFC-Based Stabilizer
Let

DVEtxy

DIExy
DvEtx

;
DvEty

DiEx
DiEy

DVBtxy
DIE2xy
DvBtx
DvBty

DiE2x

;
DVE2xy
Dvx2
Dvy2

;
DiE2y

Et jxEIE V
E; V
Et jxE2 xB IE2 V
B V
2
V
that is,
IE 1 V
Et V
E
jxE
1
IE2
Et V
B V
2
V
jxE2 xB
7:77

Bt
V
xB
Et V
B V
2 V
B
V
xE2 xB
7:78
By using Eq. (7.59), linearization of Eqs. (7.77) and (7.78) becomes the
following:
DIExy aEdc DVdc aEm DmE aEd DdE aTEEt DVEtxy
DIE2xy aE2dc DVdc aE2m DmB aE2d DdB
aTE2Et DVEtxy
7:79
aTE22 DVE2xy
DVBtxy bBtdc DVdc bBtm DmB bBtd DdB

bTBtEt DVEtxy bTBt2 DVE2xy
7:80
347
By using Eqs. (7.79) and (7.80), linearizing Eq. (4.73) gives the following:
DV_ dc cdc DVdc cmE DmE cdE DdE cmB DmB
7:81
cdB DdB cTEt DVEtxy cT2 DVE2xy

From Eq. (7.74), it can have the following:
EtI V
Et IE IE2
PEt jQEt V
E1
7:82
Hence by using Eq. (7.79), linearization of Eq. (7.82) can give the following:
DPEt cpdc DVdc cpmE DmE cpdE DdE cpmB DmB
cpdB DdB cTpEt DVEtxy cTp2 DVExy2
7:83
DQEt cqdc DVdc cqmE DmE cqdE DdE cqmB DmB

cqdB DdB cTqEt DVEtxy cTq2 DVExy2
Because
h
DVEt vVEtx0
Et0
i
vEty0
VEt0 DVEtxy ;
7:84
the state-space realization of four UPFC controllers of Eq. (7.72) can be written as
follows:
X_ UPFCAC AUPFCAC XUPFCAC BUPFCAC DVEtxy
DmE CUPFCAC XUPFCAC DUPFCAC DVEtxy Dupss
7:85
X_ UPFCDC AUPFCDC XUPFCDC BUPFCDC DVdc
7:86
DdE CUPFCDC XUPFCDC DUPFCDC DVdc

X_ UPFCP AUPFCP XUPFCP BUPFCP DPEt
DmB CUPFCP XUPFCP DUPFCP DPEt
7:87
X_ UPFCQ AUPFCQ XUPFCQ BUPFCQ DQEt
7:88
DdB CUPFCQ XUPFCAC DUPFCQ DQEt

Denote
XUPFC
DVdc
XTUPFCAC
XTUPFCDC
XTUPFCP
XTUPFCQ
T
7:89
348
By using Eqs. (7.83), (7.85), and (7.86), Eqs. (7.87) and (7.88) can be written as
X_ UPFCP AUPFCPP XUPFC BUPFCPEt DVEtxy BUPFCP2 DVE2xy BUPFCPPSS Dupss
DmB CUPFCPP XUPFC DUPFCPEt DVEtxy DUPFCP2 DVE2xy DUPFCPPSS Dupss
7:90
X_ UPFCQ AUPFCQQ XUPFC BUPFCQEt DVEtxy BUPFCQ2 DVE2xy BUPFCQPSS Dupss
DdB CUPFCQQ XUPFC DUPFCQEt DVEtxy DUPFCQ2 DVE2xy DUPFCQPSS Dupss
7:91
By substituting the second equation of Eqs. (7.85), (7.86), (7.90), and (7.91) into
Eq. (7.81), it can have the following:
DV_ dc cDCUPFC XUPFC cTDCEt DVEtxy cTDC2 DVE2xy cDCPSS Dupss 7:92
From the rst equation of Eqs. (7.85), (7.86), and (7.90)(7.92), the overall state
equation of the UPFC dynamic and control functions can be written as
X_ UPFC AUPFC XUPFC BUPFCEt DVEtxy BUPFC2 DVE2xy BUPFCpss Dupss
7:93
Let
h
DVUPFC DVTEtxy
DVTE2xy
iT
h
DIUPFC DITExy
DITE2xy
iT

X_ UPFC AUPFC XUPFC BUPFC DVUPFC BUPFCpss Dupss
DIUPFC CUPFC XUPFC DUPFC DVUPFC DUPFCpss Dupss
7:94
Similar to Eq. (6.98) and (7.42), let the linearized network equation of the
N-machine power system with the UPFC-based stabilizer installed be
3 2
Ygg
DIxy
4 DIUPFC 5 4 YUPFCg
YOg
0
2
YgUPFC
YUPFC
YOUPFC
3
DVxy
YUPFCO 54 DVUPFC 5
YO
DVO
Yg2
32
7:95
Similar to the derivation from Eqs. (7.43)(7.46), from Eqs. (7.94) and (7.95) the
open-loop, state equation of the system without the UPFC stabilizer installed can be
obtained. The output equation can be obtained by using Eqs. (7.83), (7.85), (7.86),
(7.90), and (7.91).
7.2 Design of a Shunt VSC-Based Stabilizer
7.2
349
Design of a Shunt VSC-Based Stabilizer

by Localized Phase Compensation Method
to Suppress Inter-area Line Power Oscillations
In Sect. 2.3.1.3, the phase compensation method is introduced for the design of a
PSS on the basis of the HeffronPhillips model. In Sect. 4.2.2, it is demonstrated
that when a general linearized model is used, a SSSC stabilizer can be designed by
the use of the phase compensation method. Those are examples of applying the
phase compensation method in a single-machine innite-bus power system. They
are the cases of suppressing a local-mode power oscillation along a transmission
line, where the power oscillation is closely related to the rotor motion of the
synchronous generator which is connected at one end of the transmission line.
Acceleration and deceleration of the generator absorbs and releases the active
power in responding to the power oscillation dynamically. Hence if a stabilizer (the
PSS or FACTS-based stabilizer) is designed to supply positive damping torque to
damp the oscillation of the rotor motion of the generator, it thus can effectively
suppress the local-mode power oscillation.
As it is introduced in Chaps. 5 and 6, a stabilizer in a multi-machine power
system contributes the damping torque to the electromechanical oscillation loop of
every generator. Hence, the application of the phase compensation method involves
the dynamic interactions of all generators, as it is shown in Sect. 6.2.4.2. The main
reason is that the phase compensation method has been introduced on the basis of
the damping torque analysis. In addition, like the modal analysis, the damping
torque analysis (hence the phase compensation method) needs the global information of the multi-machine power system to establish the linearized model of the
whole power system, which in practice may not always be readily available and is
difcult to be validated when the system is large and complex.
This section introduces a different phase compensation method for the design of
a VSC-based stabilizer to suppress an inter-area line power oscillation. It is
established straight on the damping analysis of the inter-area mode line power
oscillation along the transmission line where the VSC-based stabilizer locates,
rather than relying on the concept of the damping torque analysis. Its application
considers the direct contribution from the stabilizer to the damping of the inter-area
line power oscillation and does not involve in explicit consideration of interactive
dynamics of generators. Moreover, it does not need to obtain and validate the
complete power system information for applying the phase compensation method to
design the VSC-based stabilizer, because the design only requires the locally
available parameters of the transmission lines and the VSC-based unit. Hence, it is
named as the localized phase compensation method.
350
7.2.1
Localized Small-Signal Model of a VSC-Based Unit

in a Multi-machine Power System
7.2.1.1
Line Power Small-Signal Oscillation [6]
Figure 7.5 shows a VSC-based unit installed in a multi-machine power system. The
unit has a DC/DC converter behind the DC capacitor, Cdc , which can be connected
to an energy storage system. It can also be connected to a power generation plant,
such as a fuel cell (FC) or photovoltaic (PV) power plant. Without the DC/DC
converter (Idc2 0), the unit is a STATCOM. From Eq. (7.13), it can have the
following:
c mkc Vdc \c mkc Vdc \u /
V
7:96
c Vc \c and
where the modulation phase, /, is the phase difference between V

Vs Vs \u at node s, i.e. / c u. From Fig. 7.5, it can have the following:

S V
C
V

B
Vs jxsB IsB VB jXsB IAs
V
jxs
xsB
xsB
B
Vs
Vc V
jxsBIAs
xs
xs
7:97

xsB
c VB
s jxsB IAs
V
V
1 xxsBs
xs 1 xxsBs
1 xxsBs
Fig. 7.5 A VSC-based unit

in a multi-machine power
system
7:98

x As
VA
IAs
Node s
x sB
Vs
I sB
xs
Ic
Node A
Vc
m
DC/AC VSC
I dc1
Cdc
DC/DC
converter
VB
Node B
Idc2
dc
power storage or generation unit

the xy coordinate
351
VB
Va
bVB
aVc
VA
X
Hence, it can be obtained that

A jxAsIAs V
s jxAs xs xsB IAs
V
xs xsB
xsB
xs
B jxRIAs V
a
Vc
V
xs xsB
xs xsB
7:99
where
xR xAs
xs xsB
;
xs xsB
a
V
xsB
xs
B aV
c bV
B
Vc
V
xs xsB
xs xsB
From Eq. (7.99) and the phasor diagram of Fig. 7.6, the active power delivered
along the transmission line from node A to s in Fig. 7.5 can be obtained as
Va VA
VA
sin d0
bVB sin dAB akmVdc sin c
xR
xR
VA
VA
bVB sin dAB

akmVdc sinu /
xR
xR
PA
7:100
Hence, the small-signal variation of line power is given as

DPA hAB DdAB hA DVA hB DVB hu Du
hdc DVdc hm Dm hf D/
7:101
352
where
@PA
@PA
@PA
@PA
;
; hA
; hB
; hu
@dAB
@VA
@VB
@u
@PA
VA0
hdc
akm0 sinu0 /0 DVdc

@Vdc
xR
@PA VA0
hm
akVdc0 sinu0 /0
@m
xR
@PA VA0
hf
akm0 Vdc0 cosu0 /0

@/
xR
hAB
Equation (7.101) indicates that the small-signal variation of line power is directly
affected by the variation of line terminal variables, DdAB , DVA , DVB , Du and the
internal variables of the VSC-based unit, DVdc , Dm, D/:
7.2.1.2
Control Function Model of the VSC
The VSC is controlled by the DC/AC conversion to regulate active and reactive
power exchange between the VSC-based unit and the rest of the power system. Two
most popular control algorithms are the pulse width modulation (PWM) and the
pulse amplitude modulation (PAM). Figure 7.7 shows the conguration of a voltage
control type of the PWM, where m and / are the modulation ratio and phase of the
PWM, respectively.
From Fig. 7.7, the linearized model of the PWM VSC control functions can be
written as

Kac
Dm Kpac
DVs Dupss
s

7:102
Kdc
D/ Kpdc
DVdc Dupss
s
where Kpac , Kac , Kpdc , and Kdc is the gain of the AC and DC voltage proportional
and integral (PI) controllers. Obviously, if the PWM VSC controllers are not PI
ac voltage controller
u pss
Vs
+
K pac
dc voltage controller
u pss
Vdc
m
K pdc
Vdc-ref
Vs-ref
K ac
s
x ac
K dc
s
Fig. 7.7 Conguration of PWM VSC controllers (voltage control type)
x dc
353
controllers, then small-signal model of the VSC control can always be expressed by
the following general state-space representation as
sXPWM APWM XPWM bPWMs DVs bPWMdc DVdc
Dm CAC XPWM dAC DVs Dupss
7:103
D/ CDC XPWM dDC DVdc Dupss

where XPWM xac xdc T is the vector of state variables of the PWM PI controllers as shown in Eq. (7.103) and in Fig. 7.7.
Figure 7.8 shows the conguration of the voltage control type of the PAM VSC
control. From Fig. 7.8, the small-signal model of the PAM VSC control can be
written as
D/

Kpam Kac
Kppam
DVs Dupss DVdc
s
s
7:104
where Kp , Kppam , and Kpam is the gain of corresponding proportional or integral

controllers in Fig. 7.8. The state-space representation of Eq. (7.104) can be written as
sXPAM APAM XPAM bPAM DVs bpampss Dupss
D/ CPAM XPAM dDC DVdc
7:105
where XPAM xpam1 xpam2 the vector of state variables of the PAM PI controllers as shown in Eq. (7.104) and in Fig. 7.8.
The VSC controllers can also be the current control type, where the limit of
current can be easily applied for the protection of the VSC-based unit. Figure 7.9
shows the conguration of the current control type of the VSC PWM controllers.
As the difference in the control type (voltage or current) does not affect the following discussion, the voltage control type of the VSC controllers is used as the
representative in this section to demonstrate how the localized small-signal model
of the VSC-based unit can be established.
Fig. 7.8 Conguration of

PAM VSC controller (voltage
control type)
Vdc
Vs
+
Vs-ref
u pss
Kp
x pam1
K pam
K pam
s
x pam2
354
I sq0
Vs
+
K ac (s)
u pss-m
K rp (s)
Vsref
I sq
Vdc
+
I sd0
K dc (s)
u pss-f
K ap (s)
Vdcref
Isd
Fig. 7.9 Conguration of PWM VSC controllers (current control type)
7.2.1.3
Dynamic Equation of the VSC
c jxsIs V
s . Hence,
From Fig. 7.5, it can have V
isd
mkVdc sin /
;
xs
isq
Vs mkVdc cos /
xs
7:106
where subscripts d and q denote the d and q components of the corresponding

variable, respectively, in the dq coordinate of the VSC-based unit, which is shown
in Fig. 7.10. Linearization of the above equations is given as
Disd b1d DVdc b2d Dm b3d D/
7:107
Disq b1q DVdc b2q Dm b3q D/ b4q DVs
The active power exchange between the VSC-based unit and the rest of the
power system is (see Eq. (7.96) given as vcd mkVdc cos /; vcq mkVdc sin /)
Vdc Idc1 isd vcd isq vcq isd mkVdc cos / isq mkVdc sin /

the dq coordinate of
VSC-based unit
Vc
Vs
d
x
355
Hence
Idc1 isd mk cos / isq mk sin /
7:108
Thus, linearization of Eq. (7.108) can be obtained from Eq. (7.107) as

DIdc1 bd1 DVdc bd2 Dm bd3 D/ bd4 DVs
7:109
The small-signal model of a power storage or generation unit can always be

expressed by the following state-space representation:
sXP AP XP bP DVdc bvsc Duvsc
DIdc2 CP XP dP DVdc
7:110
where XP and Duvsc are the state variable vector and input of the power storage or
generation unit, respectively. In Sect. 7.3, a FC (fuel cell) power plant is used as an
example to demonstrate the procedure to derive the small-signal model of
Eq. (7.110). A similar procedure to what is demonstrated in Sect. 7.3 can be applied
to derive the model of Eq. (7.110) for the energy storage system.
Dynamic equation of the VSC is given as
1
V_ dc
Idc1 Idc2
Cdc
7:111
Hence by using Eqs. (7.109)(7.111), linearization of Eq. (7.111) can be

obtained as
sDVdc b1 DVdc b2 Dm b3 D/ b4 DVs bdc XP
7.2.1.4
7:112
Localized Small-Signal Model of the VSC-Based Unit
From Fig. 7.5 and Eq. (7.99), it can have the following:
A jxR iAsx jiAsy V
a
V
xR iAsy Va cos d0 jxR iAsx jVa sin d0
7:113
where subscripts x and y denote the x and y components of the corresponding

variable, respectively, in the xy coordinate of the power system as shown in
Fig. 7.6. In xy coordinate, Eq. (7.113) is given as
iAsx
Va sin d0
;
xR
iAsy
Va cos d0 VA
xR
7:114
356
because
A jxAsIAs V
s
V
jxAs iAsx jiAsy XAs iAsY vsx jvsy
7:115
it can be obtained from Eqs. (7.114) and (7.115) that

Va cos d0 VA
vsx VA xAs iAsy VA xAs
xR

xAs
x

x
R
As
Va cos d0
VA
xR
xR

xAs
xR xAs
bVB cos dAB akmVdc cos c

VA
xR
xR
Va sin d0
vsy xAs IAsx xAs
xR
xAs
bVB sin dAB akmVdc sin c

xR
7:116
Linearization of Eq. (7.116) gives the following:

Dvsx a01x DVdc a02x Dm a03x Dc a04x DdAB a05x DVA a06x DVB
Dvsy a01y DVdc a02y Dm a03y Dc a04y DdAB a05y DVA a06y DVB
7:117
c , as given in
s and V
However, the modulation phase, /, is the angle between V
Eq. (7.96), that is
c / u / arctan1
vsy
vsx
7:118
from Eqs. (7.117) and (7.118), it can have the following:

Dc D/ a01 DVdc a02 Dm a03 Dc a04 DdAB a05 DVA a06 DVB
7:119

Dc ag D/ ag1 DVdc ag2 Dm ag4 DdAB ag5 DVA ag6 DVB
7:120
By substituting Eq. (7.120) into (7.117), it can be obtained that

Dvsx a1x DVdc a2x Dm a3x D/ a4x DdAB a5x DVA a6x DVB
Dvsy a1y DVdc a2y Dm a3y D/ a4y DdAB a5y DVA a6y DVB
7:121
357
s vsx jvsy . Hence, the linearization

In xy coordinate of the power system, V
of Vs can be obtained from Eq. (7.121) as
DVs
vsy0
vsx0
Dvsx
Dvsy a1f DVdc a2f Dm
Vs0
Vs0
a3f D/ a4f DdAB a5f DVA a6f DVB
7:122
Finally from Eqs. (7.101), (7.102), (7.104), (7.110), (7.112), and (7.122), the
localized small-signal model of the VSC-based unit can be shown by the block
diagram in Fig. 7.11 when the VSC adopts the PWM control algorithm, and
Fig. 7.12 when the PAM is adopted by the VSC. To distinguish the stabilizer added
on the AC and DC voltage control loops of the PWM VSC (see Fig. 7.7) as well as
on the control function of the PAM VSC (see Fig. 7.8), in Figs. 7.10 and 7.11, the
stabilizing signal is denoted as Dupss m, Dupss /, and Dupss pam, respectively.
In the following section, the corresponding stabilizer is named as the PWM reactive
power stabilizer, the PWM active power stabilizer, and the PAM stabilizer.
Obviously, the establishment of the localized small-signal model of the VSC-based
unit only requires the locally available data and parameters associated with the
transmission lines and the VSC-based unit. The global information of the whole
power system is not needed.
VA
VB
K ac
s
a 2f
hA
b2
hB
a 4f
a 5f
PA
a 6f
a1f
b4
a 3f
u pss ()
hm
+
K pac +
AB
h AB
u pss (m)
Vdc
b3
1
s b1
h dc
bp
bdc
Xp
(sI - A P ) 1
hf
+
b vsc
u vsc
K pdc +
K dc
s
Fig. 7.11 Block diagram of localized small-signal model of VSC-based unit adopting the PWM
control
358
hB
hA
h AB
AB
a 4f
VA
a 5f
VB
PA
+
a 3f
a 6f
a1f
+
u pss (pam)
Kp
s
K ppam +
K pam
s
b3
b4
1
s b1
(sI - A P ) 1
b vsc
h dc
bdc
Xp
u vsc
hf
bp
Vdc
Fig. 7.12 Block diagram of localized small-signal model of VSC-based unit adopting the PAM
control
7.2.2
Design of VSC-Based Stabilizer by Localized Phase

Compensation Method
In the HeffronPhillips model of a single-machine system installed with a PSS of

Fig. 2.10, the electric torque to the electromechanical oscillation loop is contributed
to the deviation of rotor angular position Dd and PSS stabilizing signal Dupss . The
principle of superimposition of linear systems is given as
DT DTDd DTDupss
DTDd Fdelta sDd; DTDupss Fpss sDupss
7:123
where Fdelta s) and Fpss s) are the transfer functions of the forward path from Dd and
the stabilizing signal of the PSS to DT, the electric torque contribution to the
electromechanical oscillation loop, respectively. The damping torque analysis is the
359
decomposition of the electric torque on Dd Dx coordinate into a synchronizing

torque (proportional to Dd) and damping torque (proportional to Dx),
DTDd Fdelta jxs Dd Cdelta Dd Ddelta Dx
DTDupss Fpss jxs Dupss Cpss Dd Dpss Dx
7:124
Since only the damping torque contribution affects the damping of the rotor
motion, the phase compensation method is to ensure that the direct contribution
from the PSS is only the damping torque; that is, DTDupss Dpss Dx. In
Eq. (7.124), DTDd Cdelta Dd Ddelta Dx represents the effect of the variable (i.e.
Dd external to the PSS on the damping of power oscillation. It is not considered
when the PSS is designed (though it is a part of dynamic of the power system).
As it has been explained in Sect. 2.3.2.2, in fact, the damping torque analysis
converts the problem of suppression of the local-mode line power oscillation (i.e.
DPt ) into that of the damping of the rotor motion (i.e. Dd. That is why the damping
torque analysis in a multi-machine power system involves all generators. However,
if the damping of the local-mode line power oscillation, DPt , is considered directly,
from Fig. 2.10, Eqs. (7.123) and (7.124), it can have the following:
DPt DPt Dd DPt Dupss
DPt Dd K1 Dd Cdelta Dd Ddelta Dx
DPt Dupss Cpss Dd Dpss Dx
7:125
According to the damping torque analysis, only the component in DPt proportional to Dx determines the damping of line power oscillation. The phase compensation method is to ensure that the PSS only contributes the damping
component. The localized phase compensation method introduced below is based
on the establishment of Eq. (7.125) in a multi-machine power system.
According to the principle of superimposition of linear systems, from Figs. 7.11
and 7.11, it can have the following:
DPA DPA DE DPA Dupss
DPA DE FAB s
FA s FB s
Fu s Fvsc sDE
7:126
DPA Dupss Fpss sDupss

where DE DdAB DVA DVB Du Duvsc T ; Dupss is Dupss m, Dupss / or
Dupss pam; FAB s, FA s, FB s,Fu s; Fvsc s, and Fpss s is the transfer function
of the forward path from DdAB , DVA , DVB , Du, Duvsc , and Dupss to DPA in
Figs. 7.11 and 7.12, respectively.
360

T
In fact, by denoting XVSC DVdc XTPWM XTP , from Eqs. (7.101), (7.103),
(7.105), (7.110), (7.112), and (7.126), it can be obtained that
sX_ VSC AVSC XVSC BVSC DE BPSS Dupss
DPA CVSC XVSC DVSC DE dPSS Dupss
7:127
Hence comparing Eqs. (7.126) and (7.127), it can have

DPA DE CSVC sI ASVC 1 BSVC DSVC DE
DPA Dupss CSVC sI ASVC 1 BPSS dpss Dupss
7:128
7:129
That is,
FAB s
FA s
FB s Fu s
Fvsc s CSVC sI ASVC 1 BSVC DSVC

7:130
Fpss s CSVC sI ASVC 1 BPSS dpss Dupss
7:131
In Eq. (7.129), DPA Dupss is the direct contribution from a VSC-based stabilizer
to the line power variation (oscillation), DPA . Obviously, Eqs. (7.128) and (7.129)
are the extensions of Eq. (7.125) in the multi-machine power system.
Main Conclusion: If the stabilizer is designed to ensure DPA Dupss to be proportional to the time derivative of DPA , that is
DPA Dupss Dpss DP_ A Dpss [ 0
7:132
the stabilizer will supply positive damping to suppress the inter-area line power
oscillation.
The above main conclusion can be established by using the graphical explanation (similar to the introduction in Sect. 2.3.2.2) based on the linearized P d curve
and equal-area criterion as follows.
Figure 7.13 shows the linearized PA dAB curve, where dAB0 ; PA0 is the
operating point of the power system at steady state. It is assumed that the
small-signal oscillation of DPA starts from point a in Fig. 7.13 with the operating
point moving down. In dAB Dd_ AB coordinate, DPA DE can always be decomposed to be
DPA DE Cother DdAB Dother Dd_ AB
7:133
where Cother and Dother are two constants. Without affecting the establishment of the
main conclusion, it is assumed that Dother [ 0. When there is no stabilizing control
(DPA Dupss 0), DPA DPA DE Cother DdAB Dother Dd_ AB which is shown
by the dashed curve in Fig. 7.13. This is because when the operating point moves

Fig. 7.13 Graphical
explanation of the main
conclusion
361
PA(E)=CotherAB+DotherAB a
Direction
. of
PA<0
PA(E)=CotherAB
A1
PA0 g
A2
PA=PA(E)+DpssPA
f
AB2'
AB
down, Dd_ AB \0: Hence, Dother Dd_ AB \0 is added on the line DPA DE
Cother DdAB . When the operating point arrives at point f, it stops moving
(Dother Dd_ AB 0. Hence, it should be on the line DPA DPA DE Cother DdAB .
According to the equal-area criterion, the area ade is equal to that of dgf.
When the stabilizing control is added and is set according to Eq. (7.132),
Dpss DP_ A \0 (DP_ A \0) will be added on the dashed curve DPA DE
Cother DdAB Dother Dd_ AB as shown in Eq. (7.126). Hence, when the operating point
moves down, it should move below the dashed curve along the highlighted trajectory in Fig. 7.13. When the operating point stops on the line
DPA DE Cother DdAB , the area A1 is equal to area A2 at point c, where DP_ A
0 and Dother Dd_ AB 0: Obviously, dAB1 dAB0 [ dAB0 dAB2 , indicating that
extra positive damping is provided by the stabilizing control to the power oscillation. A similar analysis can be carried out to examine the case when the operating
point moves up from point c. Thus, the above main conclusion is established.
Hence, if DPA is used as the feedback signal of the VSC-based stabilizer and the
transfer function of stabilizer is constructed by a lead-lag, then it can have the
following:
Dupss Kpss
sTw 1 sT2 1 sT4

DPA
1 sTw 1 sT1 1 sT3
7:134
According to Eqs. (7.129), (7.132), and (7.134), parameters of the stabilizer can
be set at the angular oscillation frequency, xs , to satisfy the following equation
362
Fpss sKpss
sTw 1 sT2 1 sT4

Dpss s
1 sTw 1 sT1 1 sT3
7:135
with s jxs . That is, the stabilizer is set with a phase to compensate the phase of
Fpss s
s such that a pure positive damping is provided to the line power oscillation.
7.2.3
Robustness of an ESS-Based Stabilizer to Variation

of Line-Loading Conditions
7.2.3.1
Localized Small-Signal Model of an Energy Storage System [7]
In Fig. 7.5, the shunt-connected VSC-based unit can be an energy storage system
(ESS). At the steady-state operation of the power system, there is no exchange of
active power between the ESS and the rest of power system, i.e. /0 0. Hence
from Eq. (7.106), it can have isd0 0 and
mkVdc sin / k sin /0
Vdc0 Dm m0 DVdc
xs
xs
m0 kVdc0 cos /0
m0 kVdc0
Vc0
D/
D/
D/
xs
xs
xs
Vs0 Vc0
xs
Disd D
isq0
7:136
From the above equation, linearization of Eq. (7.108) can be obtained as

DIdc1 isd0 k cos /0 Dm m0 k sin /0 D/
isq0 k sin /0 Dm m0 k cos /0 D/
m0 k cos /0 Disd m0 k sin /0 Disq
Vc0
isq0 m0 kD/ m0 k
D/
xs
7:137
Hence, linearization of Eq. (7.111) for the ESS is given as

sDVdc kdc D/
DIdc2
Cdc
7:138
where kdc is a constant.

Denote the transfer function of the AC and DC voltage controllers of the VSC
adopting the PWM algorithm (see Fig. 7.7 and Eq. (7.102) where the controllers
adopt the PI control law) generally to be Kac s and Kdc s, the supplementary
363
damping control signal superimposed on the AC and DC voltage control functions,

respectively, to be upssm and upssf , that is
m m0 Kac sVsref Vs upssm
/ /0 Kdc sVdcref Vdc upssf
7:139
Because the ESS supplementary damping controller associated with upssm and
upssf is implemented through regulating, respectively, the exchange of reactive and
active power between the ESS and the rest of the power system, it is named as ESS
reactive power PSS (power system stabilizer) and active power PSS
correspondingly.
Similarly, denote the transfer functions of the voltage controllers of the VSC
adopting the PAM algorithm generally to be K1 s and K2 s (see Fig. 7.8 and
Eq. (7.104) where the controllers adopt the I and PI control law, respectively), the
added damping control signal to be upsspam , that is,
/ /0 K1 s

1
1
K2 sVsref Vs upsspam Vsref Vdc
k
k
7:140
The damping controller associated with upsspam is called the PAM PSS. By using
Eq. (7.122), linearization of Eqs. (7.138)(7.140) can be obtained as
Kac sa1f DVdc a3f D/ a4f DdAB a5f DVA a6f DVB
1 a2f Kac s
1
Dupssm
1 a2f Kac s
Dm
7:141
D/ Kdc sDVdc Dupssf

D/
K1 sK2 sa4f DdAB a5f DVA a6f DVB

k a3f K1 sK2 s
K1 sK2 s
K1 sK2 sa1f kK1 s
Dupsspam
DVdc
k a3f K1 sK2 s
k a3f K1 sK2 s
7:142
From Eqs. (7.138), (7.141), and (7.142), localized small-signal model of the ESS
is shown in Figs. 7.14 and 7.15, respectively, for the VSC adopting the PWM and
PAM, where DPA uESS hdc DVdc hm Dm hf D/; uESS DVdc Dm D/ T .
7.2.3.2
Robustness of the ESS-Based Stabilizer to the Variations

of Line-Loading Conditions
In Fig. 7.13, when the operating point moves along the linearized PA dAB curve,
sign of Dd_ AB and DP_ A is always same and Dd_ AB 0 when DP_ A 0. Hence, it can
364
7
u pss m
AB
VA

m
1
1 a 2f K ac (s)
+
K ac (s)
1 a 2f K ac (s)
a 4f
a 5f
VB
hm
a1f
a 6f
k dc
Idc2
PA (u ESS )
1
Cdc
a 3f
u pss f
1
s
Vdc
h dc
K dc (s)
hf
Fig. 7.14 Localized small-signal model of the ESS with the VSC adopting the PWM algorithm
AB
VA
VB
a 4f
PA (u ESS )
a 5f
Idc2
a 6f
K1 (s)K 2 (s)
k + a 3f K1 (s)K 2 (s)
u pss pam
hf
K1 (s)K 2 (s)
k + a 3f K1 (s)K 2 (s)
1
Cdc
k dc
1
s
Vdc
h dc
a1f K1 (s)K 2 (s) + kK1 (s)

k + a 3f K1 (s)K 2 (s)
Fig. 7.15 Localized small-signal model of the ESS with the VSC adopting the PAM algorithm
be easily concluded that in Eq. (7.126), if the decomposition of DPA Dupss in

DdAB Dd_ AB coordinate is
DPA Dupss CPSS DdAB DPSS Dd_ AB
7:143
the stabilizer associated with Dupss can suppress the line power oscillation if and
only if DPSS [ 0. The bigger DPSS [ 0 is, the more damping is provided by the
stabilizer.
Figures 7.16 and 7.17 show the block diagram of the direct damping contribution from the ESS reactive power PSS and active power PSS to the small-signal
oscillation of line power, DPA Dupssm and DPA Dupssf , respectively, when the

m
1
1 a 2f K ac (s)
u pss m
365
hm
PA (u pss m )
Fig. 7.16 Forward path of ESS reactive power PSS

m
hm
K ac (s)
1 a 2f K ac (s)
a1f
PA (u pss f )
a 3f
u pss f
1
s
k dc
Vdc
h dc
K dc (s)
hf
Fig. 7.17 Forward path of ESS active power PSS
PWM algorithm is used. In Fig. 7.16, if Dm is decomposed in DdAB Dd_ AB

coordinate to be
Dm C0pssm DdAB D0pssm Dd_ AB
7:144
from Fig. 7.15, Eqs. (7.101), (7.126), (7.143), and (7.144), it can have (/0 0)
DPSS hm D0pssm
VA
akVdc0 sin u0 D0pssm
xR
7:145
At a higher line-loading condition, u0 and hence DPSS is bigger. This means the
ESS reactive power PSS is more effective to suppress the line power oscillation at a
higher line-loading condition.
In the case of the ESS active power PSS when the PWM control is used as
shown in Fig. 7.17, if the decomposition of Dm, DVdc , and D/ in DdAB Dd_ AB
coordinate, respectively, is
Dm Cpssfm DdAB Dpssfm Dd_ AB
DVdc Cpssfdc DdAB Dpssfdc Dd_ AB
D/ Cpssff DdAB Dpssff Dd_ AB
7:146
366
from the above decomposition, Fig. 7.17, Eqs. (7.101), (7.126), (7.143) and
(7.144), then it can have (/0 0)
DPSS hm Dpssfm hdc Dpssfm hf Dpssff
VA
akVdc0 sin u0 Dpssfm m0 sin u0 Dpssfdc m0 Vdc0 cos u0 Dpssff

xR
7:147
When the line-loading condition increases, sin u0 increases but cos u0 decreases
in Eq. (7.147). Hence, DPSS does not change as much as it does in the case of the
ESS reactive power PSS. This means that the robustness of the ESS active power
PSS to the variations of line-loading conditions is expected to be better than that of
the ESS reactive power PSS. The same conclusion can be drawn from Fig. 7.18 for
the ESS PAM PSS when the PAM control is used.
The same conclusion has been established in Sect. 4.1.2.2 in the case of a
single-machine innite-bus power system. Function of the ESS reactive power
stabilizer involves the regulation of magnitude of the terminal voltage of the VSC.
Hence, in essence it works via controlling the exchange of reactive power between
the ESS and the rest of the power system. Both the ESS active power stabilizer and
the PAM stabilizer function through the regulation of phase of the terminal voltage
of the ESS, that is, the exchange of the active power between the ESS and the rest
of the power system. Therefore, the ESS active power stabilizer and PAM stabilizer
suppress the line power oscillation directly by absorbing active power from or
injecting it into the power system when they see the excess or lack of active
power delivered along the transmission line. Hence, their capability to damp the
power oscillation is less affected by the change of system loading conditions.
The ESS reactive power stabilizer affects indirectly the variation of active power
delivered along the transmission line through the regulation of injection or
absorbing of reactive power from the ESS. At a lower loading condition, regulation
of reactive power exchange is related less closely with the variation of line active
power. Hence, the stabilizer is less effective in damping the power oscillation.
PA (u pss pam )
hf
u pss pam
K1 (s)K 2 (s)
k + a 3f K1 (s)K 2 (s)
k dc
1
s
Vdc
a1f K1 (s)K 2 (s) + kK1 (s)

k + a 3f K1 (s)K 2 (s)
Fig. 7.18 Forward path of ESS PAM PSS
h dc
7.3 An Example of a Multi-machine Power System
7.3
367
An Example of Multi-machine Power System

with a Grid-Connected FC Power Plan
Fuel cell (FC) power generation is one of the most promising technologies of clean
energy for the future. Among several types of the FC, the high-temperature solid
oxide fuel cell (SOFC) is considered to be the most attractive FC technology for
grid-connected power generation. It is foreseen that, in future power systems,
gird-connected high-capacity SOFC power plants could play certain important role
as one of the main clean power generation sources.
This section presents an example of a multi-machine power system with a SOFC
power plant which is integrated into the power system via a shunt VSC. The
objective is to demonstrate how the linearized mathematical model of the power
system with the shunt VSC-based SOFC power plant can be established for power
system small-signal stability analysis and control. It shows how the localized
small-signal model of the SOFC power plant can be developed and a stabilizer
attached to the VSC of the SOFC power plant can be effectively designed by using
the localized phase compensation method introduced in Sect. 7.2.
7.3.1

System with a Grid-Connected FC Power Plant
7.3.1.1
Dynamic Model of the SOFC Power Plant
Power generation by a SOFC is the conversion of chemical energy in hydrogen

(H2) and oxygen (O2) into electric energy through the following basic reaction at
two nodes of the SOFC:
H2 O ! H2 O 2e
O2
2e ! O
Cathode:
2
Anode:
DC current Ifc is generated from the releasing of electrons (2e) at the anode due
to the above chemical reaction. A dynamic model of the SOFC power plant suitable
for use in power system studies is shown in Fig. 7.19. In Fig. 7.19, electrical
dynamic describes the chemical reaction to restore the charge that has been quickly
drained by the load. This is modelled by a rst-order transfer function. To ensure
the FC operation within the safe operating area, current control reference, Ifcref, is
limited by the following boundaries:
Ifcrefmax
Umax
Umin
N0
q
; Ifcrefmin
q
; Kr
2Kr h2in
2Kr h2in
4F
368
Ifc-ref-max
electrical
dynamic
DC/DC converter
dc
DC/AC converter
m
Idc2
Pfc-ref
I fc-ref =
I fc =
Pfc-ref
Vfc
Idc1
Ic
I fc
DC/DC
1
I fc-ref
1+Te s
Idc
Vdc
AC
grid
Cdc
Vc
fuel reaction
I fc-ref-min
q h2-in
2K r 1
q h2-in =
Ifc
U opt 1+Tf s
fuel
processor
Vfc
p h2 =
1
1
(q h2-in -2K r I fc )
K h2 1+Th2 s
q h2-in
p o2 =
1
1
(q o2-in -K r I fc )
K o2 1+To2 s
1
rho
p h2o =
q o2-in
Ifc
1
1
2K r I fc
K h2o 1+Th2o s
p h2
Vfc = N 0 [E 0 +
p O2
RT p h 2 p02
ln(
2F
p h 2o
p h2O
0.5
)] - rI fc
Nernst equation
Fig. 7.19 Dynamic model of a grid-connected SOFC power plant
where Umax and Umin are the maximum and minimum fuel unitization, respectively,
N0 the number of cells in series in the FC stack, qh2in the hydrogen input flow rate,
and F the Faraday constant.
Fuel processor depicts the dynamics of fuel supply which is a process to reform
the natural gas supply to the hydrogen-rich fuel. This can be simply represented by
a rst-order lag model of time constant Tf, where Uopt is the optimal fuel utilization.
The chemical process of fuel reaction inside the FC stack is described by three
rst-order transfer functions for hydrogen, oxygen, and water, respectively, where
rho is the ratio of hydrogen to oxygen, Kh2, Ko2, and Kh2o the valve molar constant,
Th2, Ko2, and Th2o the time constant and ph2, po2, and ph2o the partial pressure for
hydrogen, oxygen, and water flow, respectively. The internal EMF generated by the
SOFC stack is calculated by the well-known Nernst equation, where E0 is the ideal
standard potential, R the universal gas constant, T the absolute temperature, and r
the ohmic loss.
Grid connection of the SOFC power plant can be implemented by a two-stage
conguration of power electronics interface as shown in Fig. 7.19. In order to avoid
the underused or overused fuel, the DC/DC converter can be controlled to ensure a
constant current output from the fuel cell stack. That is,
dc dc0 Tdc sIfcref Ifc
7:148
where dc is the duty cycle and Tdc s the transfer function of the FC current
controller.
369
VA
x As
Node s
IAs
Vs
Node A
x sB
VB
IsB
Ic
xs
Node B
Vc
SOFC
power plant
Fig. 7.20 A multi-machine power system with a SOFC power plant connected at node s
Figure 7.20 shows the SOFC power plant connected to a multi-machine power
system at node s via a VSC. Voltage at the AC terminal of the DC/AC converter of
the SOFC power plant is (see Eq. (7.96)) given as
c mkVdc \c mkVdc \u /
V
7:149
If the DC/AC converter is controlled by the PWM control, m and / are the
modulation ratio and phase of the PWM, respectively. If the control algorithm of
the DC/AC converter is the PAM, then m = 1.
From Eq. (7.108), it can have the following:
Idc1 isd mk cos / isq mk sin /
7:150
where subscripts d and q denote the d and q components of the corresponding

variable, respectively, in the dq coordinate of the SOFC power plant as shown in
Fig. 7.20. Because
Pfc Ifc Vfc Idc2 Vdc ;
Vdc
Vfc
;
1 dc
7:151
it can have Idc2 1 dc Ifc such that the dynamics of the DC/AC converter can be
expressed as
1
V_ dc
Idc1 Idc2
Cdc
1
isd mk cos / isq mk sin / 1 dc Ifc

Cdc
7:152
From the point of view of power system operation, node s can be treated as a PV
bus. Active power output from the SOFC power plant is kept to be pfc-ref because of
the current controller implemented in the DC/DC converter. Voltage control
scheme applied by the DC/AC converter will be very useful as far as the power
370
system operation control is concerned. This can be achieved by either the PWM or
the PAM algorithm as shown in Fig. 7.22 (see Figs. 7.7 and 7.8), where Kac s and
Kdc s are the two transfer functions of the PWM- and PAM-controlled DC/AC
converter, and upss is a supplementary damping control signal to improve power
system stability.
7.3.1.2
Linearized Model of the SOFC Power Plant
Similar to Eq. (7.106), from Fig. 7.20, it can have the following:
isd
mkVdc sin /
;
xs
isq
Vs mkVdc cos /
xs
7:153
s vsx jvsy . Hence,

In the common xy coordinate of the power system, V
linearization of Vs can be obtained as
DVs vTs0 DVsxy
7:154

h
i
Dvsx
v
where vTs0 vVsx0 Vsy0 ; DVsxy
: From Eqs. (7.153) and (7.154), it can
s0
s0
Dvsy
have the following:
Disd b1d DVdc b2d Dm b3d D/
Disq b1q DVdc b2q Dm b3q D/ bT4q DVsxy
7:155
If the linearization of Eq. (7.148) is Ddc Tdc sDIfc and the state-space
realization of transfer function Tdc s is Ad ; Bd ; Cd ; Dd , then it can have the
following:
X_ d Ad Xd Bd DIfc
Ddc Cd Xd Dd DIfc
7:156
If the state-space realization of transfer function Kac s and Kdc s is

AAC ; BAC ; CAC ; DAC and ADC ; BDC ; CDC ; DDC , respectively, then from
Fig. 7.22 it can have that for the PWM control
X_ AC AAC XAC BAC DVsxy
Dm CAC XAC DAC DVsxy Dupss
D/ CDC XDC DDC DVdc Dupss
7:157
7:158
371
and for the PAM control

X_ AC AAC XAC BAC DVsxy bpampss Dupss
D/ CDC XDC DDC CAC XAC DDC DAC DVsxy DDC DVdc dpampss Dupss
7:159
By using Eqs. (7.155)(7.159), linearization of Eq. (7.152) can be obtained as
DV_ dc a1 Vdc a2 DIfc aT3 Xd aT4 XAC
aT5 XDC aT6 DVsxy a7 Dupss
7:160
where Dupss can be Dumpss , Du/pss or Dupampss with different coefcient a7.
Linearization of dynamic model of the SOFC power plant of Fig. 7.19 can be
obtained as
Pfcref0
1
DIfcref
DVfc ; DIfc
1 Te s
V2fc0
2Kr
1
1
DIfc ; Dqo2in
Dqh2in
Dq
rho h2in
Uopt 1 Tf s
1
1
Dqh2in 2Kr DIfc
Dph2
Kh2 1 Th2 s
1
1
Dqo2in Kr DIfc
Dpo2
Ko2 1 To2 s
1
1
2Kr DIfc
Dph2o
Kh2o 1 Th2o s
DVfc c1 Dph2 c2 Dp02 c3 Dph2o c4 DIfc
DIfcref
7:161
Rearranging Eq. (7.161) can give the following:

DI_fc d1 Dph2 d2 Dp02 d3 Dph2o d4 DIfc
Dq_ h2in d5 Dqh2in d6 DIfc
Dp_ h2 d7 Dqh2in d8 Dph2 d9 DIfc
7:162
Dp_ o2 d10 Dqh2in d11 Dpo2 d12 DIfc

Dph2o d13 Dph2o d14 DIfc
From Eqs. (7.156)(7.160) and (7.162), the state-space equation of the SOFC
power plant can be obtained as
X_ SOFC ASOFC XSOFC BSOFC DVsxy BPSS Dupss
7:163
372

the dq coordinate of the
SOFC power plant
Vc
Vs
where

XSOFC DVdc
DIfc
Dqh2in
Dph2
Dpo2
Dph2o
XTd
XTAC
XTDC
T
v
Since u ac tan1 vsysx , it can have the following:
Du aT0 DVsxy
h v
sy0
where aT0 v2 v2
sx0
sy0
7:164
i
vsx0
v2sx0 v2sy0 . From Fig. 7.21, it can have the following:
isx
isy
cos u
sin u
sin u
cos u
isd
isq

7:165
Linearization of Eq. (7.165) is given as

Disx
Disy
cos u0
sin u0
sin u0
cos u0

Disd
sin u0
Disq
cos u0
cos u0
sin u0

isd0
Du
isq0
7:166

DIsxy b1 DVdc b2 Dm b3 D/ bT4 DVsxy
7:167

Disx
where DIsxy
. By using Eqs. (7.157)(7.159) and (7.163), Eq. (7.167) can
Disy
be expressed as
DIsxy CSOFC XSOFC DSOFC DVsxy DPSS Dupss
7.3.1.3
7:168
Linearized Model of Multi-machine Power System

with SOFC Power Plant
Similar to the derivation in Sect. 6.1, it is assumed that there are N generators and
M nodes in the multi-machine power system with the gird-connected SOFC power
373
plant. In the common xy coordinate of the power system, node voltage and current
are denoted as vxj jvyj and ixj jiyj , j 1; 2; . . .; M, respectively. Without the loss
of generality, it can be assumed that node 1 to N is generator node and node s in
Fig. 7.20 is node (N + 1). Dening (see Eq. (5.51))
DVxy Dvgx1
DIxy Dix1
Dvgy1
Diy1
Dvgx2
Dix2
Dvgy2
Diy2

DvgyN T
DvgxN
DixN
DiyN T
Linearized model of generators is (see Eq. (6.33)) given as

X_ gp Agp Xgp Bgp DVxy
DIxy Cgp Xgp Dgp DVxy
7:169
where Xgp is the state variable vector of generator dynamics, including state
variables of the AVRs and the PSSs. Denote the voltage at other nodes by the
following vector:

DVoxy DVN 2x
DVN 2y
DVMx
DVMx
T
The network equations can be established as

3 2
Ygg
DIxy
4 DIsxy 5 4 Ysg
Yog
0
2
32
3
DVxy
Ygo
Yso 54 DVsxy 5
Yoo
DVoxy
Ygs
Yss
Yos
7:170
Substituting Eqs. (7.163), (7.168), and (7.169) into Eq. (7.170), it can have the
following:
3 2
Ygg Dgp
Cgp Xgp
4 CSOFC XSOFC DPSS Dupss 5 4 Ysg
Yog
0
2
32
3
DVxy
Ygo
Yso 54 DVsxy 5
Y2o
DVoxy
Ygs
Yss DSOFC
Yos
7:171
Equation (7.171) can give the following:
DVoxy Y1
oo Yog
DVxy
DVsxy
Y1
gsxy
where Ygsxy
Yos
DVxy
DVsxy
Cgp
CSOFC
Ygg Dgp
Ysg
Xgp
XSOFC
0
DPSS

Dupss

Ygs
Ygo 1

Yoo Yog
Yss DSOFC
Yso
Yos
7:172
374
Arranging the rst equation in Eqs. (7.169) and (7.163) together, it can have the
following:
"
X_ gp
X_ SOFC
Agp
Xgp
0
ASOFC XSOFC

DVxy
0
Bgp
0
Dupss
0 BSOFC DVsxy
BPSS
7:173
Substituting Eq. (7.172) into (7.173), system state equation can be obtained as
_ AX BDupss
X
7:174
where

Xgp
X
XSOFC

Agp
0
0
Cgp
Bgp
1
A
Ygsxy
0
ASOFC
0 BSOFC
0

0
0
Bgp
0
Y1
B
0 BSOFC gsxy DPSS
BPSS
0
CSOFC

;
System small-signal stability can be improved by the supplementary stabilizing

control superimposed on the normal control functions of the DC/AC converter (see
Fig. 7.22). If the feedback signal of the stabilizer is the deviation of active power
delivered along the transmission line from node A to s in Fig. 7.20, DPA , then it can
have the following:
ac voltage controller
u pss
Vs
+
m'
K ac (s)
dc voltage controller
u pss
Vdc
m
K dc (s)
'
Vdc-ref
Vs-ref
PWM
Vdc
Vs
+
Vs-ref
K ac (s)
K dc (s)
u pss
PAM
Fig. 7.22 Control of DC/AC converter either by the PWM or by the PAM algorithm
375

VA Vs
PA Re
VA RegAs jbAs VA VA Vs VA
ZAs

gAs v2Ax v2Ay bAs vAx vsy vAy vsx
From the above equation, it can have the following:

DPA 2gAs vAx0 bAs vsy0
2gAs vAy0 bAs vsx0

bAs vAx0
bAs vAy0
Dvsx
Dvsy
DvAx
DvAy
ps DVsxy p0s DVoxy
7:175
1
DPA 0 ps p0s Y1
oo Yog Yos Ygsxy

Cgp
0
Xgp
0
Dupss
DPSS
0
CSOFC XSOFC
CX DDupss
7:176
Equation (7.176) is the output equation of state-space representation of the power

system. By denoting y DPA , the linearized model of the multi-machine power
system with the gird-connected SOFC power plant is obtained as (Eqs. (7.174) and
(7.176))
X_ AX BDupss
y CX DDupss
7:177
This is the linearized mathematical model of the whole power system. It can be
used for system small-signal stability analysis and design of the stabilizer attached
to the VSC of the SOFC power plant.
7.3.2
Design of a Stabilizer Attached to the VSC

of FC Power Plant by Localized Phase
Compensation Method
7.3.2.1
Localized Design of the Stabilizer
Based on the linearized model of Eq. (7.177), power system small-signal stability
can be examined by modal analysis, i.e. computation of system oscillation modes
376
and associated eigenvectors. For the design of the supplementary stabilizer, various
methods proposed for the design of the PSS or FACTS stabilizers introduced in
Chap. 5 and 6 can be used. However, even the stabilizer is to suppress a locally
observed oscillation, such as the power oscillation along the transmission line from
node A to s in Fig. 7.20, and the mathematical model of the whole power system of
Eq. (7.177) needs to be established. This increases signicantly not only the
complicity of the design, but also the possibility of error in obtaining system model
of Eq. (7.177), as it is difcult to gain and validate the accurate system parameters
and operational data when the scale of the power system is large. In Sect. 7.2,
a much simpler method to design the stabilizer by using a localized linearized
model of the VSC-based unit is introduced. The stabilizer is designed by a localized
phase compensation method to suppress the power oscillation along the transmission line where the VSC-based unit is installed. In this section, a stabilizer attached
to the VSC of the grid-connected SOFC power plant is used as an example to
demonstrate the application of the localized phase compensation method as follows.
By substituting Eq. (7.121) into (7.163), the following localized state-space
equation of the SOFC power plant dynamics can be obtained as
sX_ SOFC ALSOFC XSOFC BESOFC DE BPSS Dupss
7:178
By using Eqs. (7.157)(7.159), (7.164), and (7.121), Eq. (7.101) can be written
as
DPA CLSOFC XSOFC CESOFC DE dSOFC Dupss
7:179

DPA DPA DE DPA Dupss
7:180
where
DPA DE CLSOFC sI ALSOFC 1 BESOFC CESOFC DE
DPA Dupss CLSOFC sI ALSOFC 1 BPSS dSOFC Dupss
DPA Dupss is the direct contribution from the stabilizer to the variation (oscillation) of line active power, DPA . If DPA is used as the feedback signal of the
stabilizer, then the transfer function of stabilizer is constructed by a lead-lag
network
Dupss Kpss
sTw 1 sT2 1 sT4

DPA
1 sTw 1 sT1 1 sT3
7:181
377
Denote
FSOFC s CLSOFC sI ALSOFC 1 BPSS dSOFC
7:182
According to Sect. 7.2.2, parameters of the stabilizer can be set at the angular
oscillation frequency, xs , to satisfy the following equation:
FSOFC sKpss
sTw 1 sT2 1 sT4

Dpss s
1 sTw 1 sT1 1 sT3
7:183
with s jxs .
The stabilizer will supply positive damping to suppress the line power oscillation
effectively.
7.3.2.2
An Example
Figure 7.23 shows the conguration of a four-machine two-area power system

integrated with a SOFC power plant. This is the modied example power system
used in Sect. 6.3. The system has a poorly damped inter-area oscillation mode.
Parameters of the SOFC power plant are as follows:
T 1273; F 96487; R 8:314; E0 1:18; N0 384; Kr 0:966e6;
Umax 0:9; Umin 0:8; Uopt 0:85; Kh2 8:43e4; Kh2o 2:81e4;
Ko2 2:52e3; Th2 26:1; Th2o 78:3; To2 2:91;
r 0:126; Tf 5; Te 0:08; rho 1:145; Cdc 1:0
XT2
G2
L5
PA
G1
XT1
X12
VA
X26
X56
5
X45
X34
VB
Vs
XT3
L2
Is
Xs
VC
G4
SOFC power plant
Fig. 7.23 A power system integrated with a PV generation plant
XT4
G3
378
To enhance the damping of the inter-area oscillation mode, a stabilizer can be

added onto the DC/AC VSC of the SOFC power plant as discussed in the above
section. For the purpose of demonstration, the following three cases are studied:
1. The PWM is adopted and the stabilizer is superimposed on the AC voltage
controller (see Fig. 7.22). This is the PWM reactive power PSS.
2. The stabilizer is added on the DC voltage controller (see Fig. 7.22) of the PWM.
This is the PWM active power PSS.
3. The PAM algorithm is used (see Fig. 7.22). The added stabilizer is named the
PAM stabilizer.
Three types of stabilizers are designed at the line-loading condition
PA0 = 500 MW by using the method introduced in Sect. 7.2.2 based on the localized
model of the SOFC power plant of Eqs. (7.178) and (7.179). To conrm the
effectiveness of the stabilizers designed, results of modal analysis obtained by using
the complete mathematical model of the power system of Eq. (7.177) are given in
Table 7.1 and simulation results using the nonlinear mathematical model of the
example power system are shown in Fig. 7.24. In the nonlinear simulation, the
example power system is subject to a disturbance of 1 % increase of mechanical
power input to G4 for 100 ms at 0.2 s of simulation. From Table 7.1 and Fig. 7.24,
it can be seen that the stabilizers are designed successfully by using the localized
phase compensation method to effectively suppress the inter-area power oscillation.
Three types of stabilizers are tested at a lower line-loading condition of the
example power system with PA0 = 300 MW. Results of modal analysis are given in
Table 7.1. Figure 7.25 presents the results of nonlinear simulation (subject to the
same disturbance as that in Fig. 7.24). From Table 7.1 and Fig. 7.25, it can be seen
that at the lower line-loading condition, PWM reactive power stabilizer designed at
the higher line-loading condition with PA0 = 500 MW cannot suppress the inter-area
oscillation effectively. The robustness of the PWM active power and PAM stabilizer is better as they can still function to damp the inter-area power oscillation
when the line-loading conditions of the power system change.
Similar to the case discussed in Sect. 7.2.3 about the ESS-based stabilizer, the
PWM active power and PAM stabilizer function through regulating the phase of
terminal voltage of the SOFC power plant, that is, the exchange of active power
between the SOFC power plant and the rest of the power system. Therefore, they
suppress the inter-area power oscillation directly by absorbing active power from or
injecting it into the power system when they see the excess or lack of active
Table 7.1 Inter-area oscillation mode

With PWM reactive power stabilizer
With PWM active power stabilizer
With PAM stabilizer
PA0 = 500 MW
PA0 = 300 MW
0.103
0.355
0.405
0.529
0.883
0.221
0.320
0.398
j3.471
j3.169
j3.365
j3.363
j3.602
j3.331
j3.454
j3.419

PA(100MW)
379

With PAM stabilizer
Time (second)
Fig. 7.24 Results of nonlinear simulation when the system is subject to a small disturbance
(PA0 = 500 MW)
PA(100MW)

With PAM stabilizer
Time (second)
Fig. 7.25 Results of nonlinear simulation when the system is subject to a small disturbance
(PA0 = 300 MW)
380
power delivered along the transmission line. Hence, their capability to damp the
power oscillation is less affected by the change of system loading conditions.
Function of the PWM reactive power stabilizer involves the regulation of magnitude of terminal voltage of the SOFC power plant. Hence, in essence it works via
controlling the exchange of reactive power between the SOFC power plant and the
rest of the power system. The stabilizer affects indirectly the variation of active
power delivered along the transmission line through the regulation of injection into
or absorbing of reactive power from the SOFC power plant. At a lower line-loading
condition, regulation of reactive power exchange is related less closely with the
variation of line active power. Hence, the stabilizer is less effective.
7.4
Damping of Multi-mode Oscillations by Multiple

Stabilizers Attached to a Single UPFC
To suppress power system multi-mode oscillations, the commonly used schemes

are to deploy multiple decentralized PSS or/and FACTS stabilizers. Usually, a
stabilizer is designed to improve the damping of one selected oscillation mode and
hence damping of multi-mode oscillations is achieved by multiple stabilizers
jointly. Under these schemes, parameters of multiple stabilizers are set either
individually or simultaneously. In engineering practice nowadays, strategy of
individual setting is often used because the setting can be easily tuned in eld. In
many cases, it is effective and acceptable for the multiple stabilizers to provide
reasonable damping to multi-mode oscillations. Extensive research in the last three
decades has demonstrated that the coordinated design to simultaneously set multiple stabilizers can provide better results to effectively suppress power system
multi-mode oscillations as introduced in Sects. 5.2.2 and 6.2.4.
However, from the point of view of engineering practice, at the stage of eld
tuning after parameters of multiple stabilizers are set in coordination, it could be a
difcult and challenging task to nally implement the schemes of simultaneous
setting of multiple stabilizers. This is because often parameters of stabilizers set in
coordination need to be further tuned carefully in the eld to provide required
control performance. However, those stabilizers are often at different geographical
locations in a power system. Cross-location and simultaneous eld tuning of
parameters of multiple stabilizers is practically very difcult. In this case, if
parameters of stabilizers are tuned in eld one by one in a sequence, the overall
strategy of deploying multiple stabilizers will be the combination of simultaneous
setting and individual eld tuning. It is unlikely that the optimality of stabilizers
design by setting their parameters simultaneously can be achieved eventually by
implementing individual parameters tuning in the eld.
Hence, this section demonstrates a scheme of attaching multiple stabilizers to a
single UPFC to damp power system multi-mode oscillations. The method takes the
full advantage of the UPFC, which can perform multiple control functions of active,
7.4 Damping of Multi-mode Oscillations by Multiple Stabilizers
381
reactive power, or/and voltage regulation. Multiple stabilizers are superimposed on

different control functions of a single multifunctional UPFC. Each stabilizer is set to
specially meet the need of damping improvement of one selected oscillation mode
and multiple stabilizers are designed in coordination with damp multi-mode
oscillations. This arrangement of multiple stabilizers is as same as that of conventional schemes of coordinated design of multiple stabilizers introduced in
Sects. 5.2.2 and 6.2.4. Hence, the optimal design to suppress multi-mode oscillation
can be achieved. Moreover, being attached to one UPFC, multiple stabilizers are at
a same geographical location in the power system. Therefore, problem of
cross-location eld tuning of stabilizers parameters is effectively avoided.
7.4.1
Coordinated Design of Multiple Stabilizers

by Articial Fish Swarm Algorithm (AFSA) [8]
Methods of optimal coordinated design of multiple stabilizers can be classied

under two categories: mathematical optimization and non-mathematical optimization. In Sect. 5.2.2, a representative mathematical method based on the nonlinear
programming is introduced. One of the most well-known methods of
non-mathematical optimization is the genetic algorithm (GA) which has been tested
for the coordinated design of multiple stabilizers in power systems. In this section, a
different method of non-mathematical optimizationarticial sh swarm algorithm
(AFSA)is presented for the design of multiple stabilizers attached to a single
UPFC to suppress multi-mode power system oscillation as follows.
7.4.1.1
Objective Function
The coordinated design of multiple stabilizers is based on the following objective

function
J fx minf1i ; i 1; 2; . . .; mg
7:184
where 1i is the damping of the ith oscillation mode and m the number of oscillation
modes to be damped by m stabilizers. Hence, the objective function J is the
damping of the oscillation mode of the power system that is of the least damping.
The coordinated design is to search for the maximum of J = f(x).
Transfer function of the jth stabilizer superimposed on a selected primary control
function of the UPFC is given as
Kj
sTwj 1 sT2j 1 sT4j

1 sTwj 1 sT1j 1 sT3j
Parameters of the stabilizer, T2j, T4j, and TWj are predetermined, and T1j, T3j, and
gain Kj are to be set by using the AFSA. In the parameter space of J, parameters of
382
the jth stabilizer involved in the coordinated design are subject to the following
constrains:
Kjmin Kj Kjmax ;
T1jmin T1j T1jmax ; T3jmin T3j T3jmax
The AFSA is an algorithm based on the population evolution similar to the GA.
The basic argument of the AFSA is that in a water area, the place where there are
most sh is where there is most food. The optimal search imitates the behaviours of
sh to seek food. In the n-dimensional parameter space of the objective function
J fx dened by Eq. (7.184), the algorithm starts with constructing k articial
sh and the position of the ith sh is denoted by vector xi xi1 xi2 xin .
The amount of food at
present position is denoted by Yi fxi . Distance between
articial sh is dij xi xj and the visual distance of articial sh is denoted by
Visual. Crowding index is CI and maximal moving step is Step. The AFSA is
implemented by three main behaviours of articial sh as introduced in the following section.
7.4.1.2
The Articial Fish Swarm Algorithm
Preying behaviour
It is assumed that the present position of articial sh is xi t and the amount of
food at the position is Yi t. Within the visual distance of articial sh, i.e.
dij \Visual, a new position xi t 1 is randomly chosen and Yi t 1 is calculated.
If Yi t 1 [ Yi t, then the articial sh moves one step forwards. Otherwise, a
new position is randomly chosen.
Swarming behaviour
This is to count the number of sh within the visual distance, nf and to calculate the
central position xc of this group of sh. The amount of food at the central position is
Yc t. If Yc t=nf [ CIYi t, this means that there is more food at the central
position and the position is not very crowded. Hence, the articial sh moves one
step towards the central position. Otherwise, it performs preying.
Following behaviour
If within the visual distance of an articial sh, there is most amount of food
Ymax t at position xmax . Using xmax as the centre, the number of sh nmax within its
visual distance (dij \Visual) is counted. If Ymax t=nmax [ CIYi t, then there is
more food at xmax and it is not very crowded there. The articial sh moves one
step towards xmax . Otherwise, it performs preying.
In the AFSA, preying behaviour guarantees the convergence of the algorithm.
The introduction of crowding index CI is to prevent that a local optima is found as
383
AFSA Initialization
(number of artificial fish k and
their initial position)
Calculation of amount of food
Swarming movement of
artificial fish
Following movement of
artificial fish
NO
More food found?
YES
NO
Preying movement of
artificial fish
Objective
function satisfied?
(amount of food found
satisfied?)
YES
STOP
Fig. 7.26 Flow chart of the AFSA
the nal solution. The swarming behaviours are to ensure the stability of the algorithm and global optimization. Following behaviours are to accelerate the convergence of the algorithm towards the global solution. A flow chart of implementing the
AFSA is shown in Fig. 7.26. The AFSA essentially is a random global optimization
algorithm. It is of many attractive features of the genetic algorithm, but it does not
need to perform crossover and mutation. Hence, it is relatively easier and simpler.
7.4.2
Examples
7.4.2.1
An Example of Damping Two-Mode Oscillations by Stabilizers

Added on a UPFC
The 4-machine 2-area power system of Fig. 7.24 is used as an example, where the
SOFC power plant is replaced by a UPFC and G1 and G2 are not installed with the
384
PSS. The UPFC employs four control functions as dened by Eq. (7.71). In the
system, there are three oscillation modes. One local mode related to G3 and G4 is
well damped. An inter-area mode (G1 and G2 against G3 and 4) and a local mode
(G1 against G2) are poorly damped. Two stabilizers are superimposed on two
normal control functions of the UPFC to suppress the poorly damped local-mode
and inter-area mode oscillation. To select the control functions to be attached with a
stabilizer, the controllability index is calculated and presented in Table 7.2. From
Table 7.2, it can be seen that in order to suppress two-mode oscillations, the rst
stabilizer should be added to the function of the AC voltage regulation to damp the
inter-area oscillation mode and the second stabilizer to DC voltage regulation to
damp the local oscillation mode.
Two stabilizers superimposed on the functions of the AC and DC voltage regulation of the UPFC are designed in coordination based on the linearized model of
the power system presented in Sect. 7.2. Feedback signal of the rst stabilizer is the
deviation of active power delivered from node 78 and that of the second stabilizer
is the difference between the deviation of active power from node 57 and that from
node 27. Results of coordinated design of two stabilizers by using the AFSA are
given in Table 7.3. From the third, fourth, and fth rows of Table 7.4, it can be seen
that two stabilizers can enhance the damping of two oscillation modes effectively.
Table 7.2 Results of controllability computation of the UPFC normal control functions
Control function to attach stabilizer
Inter-area mode
Local mode
Active power regulation

Reactive power regulation
AC voltage regulation
DC voltage regulation
3.77
5.85
49.29
138.26
0.10
0.11
1.41
7.43
Table 7.3 Results of coordinated design (T2 = T4 = 0.05, Tw = 5)

Parameters
The rst stabilizer
The second stabilizer
T1
T3
K
0.01
0.12
0.25
0.37
0.07
28.36
Table 7.4 Oscillation modes in the example power system with the UPFC
With the rst stabilizer
With the second stabilizer
With two stabilizers
Inter-area mode
Local mode
0.12062 j3.4785
0.70652 j3.0758
0.05775 j3.3794
0.6812 j3.0858
0.20745 j5.8104
0.21014 j5.8047
1.1759 j5.7957
1.2403 j5.6547

when the rst stabilizer is
installed. a Inter-area mode
oscillation and b local-mode
oscillation
385
(a)
0.46
delta1-3
Without stabilizer
With the first stabilizer
0.44
0.42
0.4
0.38
0.36
0.34
0.32
0.3
0.28
10
t/s
(b)
0.155
delta1-2
Without stabilizer
With the first stabilizer
0.15
0.145
0.14
0.135
0.13
0.125
0.12
0.115
10
t/s
Figures 7.27 and 7.28 present the results of nonlinear simulation when the
example power system is subject to a three-phase to-earth fault at node 8 for
100 ms. Figure 7.27 is that when only the rst stabilizer is installed. Because the
rst stabilizer is set to damp inter-area mode, in Fig. 7.27 the local-mode oscillation
is not suppressed effectively. From Fig. 7.28, it can be seen that with two stabilizers
being superimposed on the AC and DC voltage regulator of the UPFC, respectively,
two-mode oscillations of the example power system are damped successfully.
Hence, this example demonstrates that it is feasible to suppress multi-mode oscillations by arranging two stabilizers on a single UPFC.
386

when two stabilizers are
installed. a Inter-area mode
oscillation and b local-mode
oscillation
(a)
0.46
Without stabilizer
delta1-3
0.44
0.42
0.4
0.38
0.36
0.34
0.32
0.3
0.28
0
10
t/s
(b)
0.155
delta1-2
Without stabilizer
0.15
0.145
0.14
0.135
0.13
0.125
0.12
0.115
10
t/s
7.4.2.2
An Example of Damping Three-Mode Oscillations

by Stabilizers Added on a Multi-terminal UPFC
In order to show that the example presented in the above section is not the unique
case, the conguration of the example power system of Fig. 7.24 is modied to
form a 6-machine power system as shown in Fig. 7.29. In this example power
system, there are three weakly damped oscillation modes as shown in the second
row of Table 7.5. They are: (1) the rst inter-area oscillation mode (rst mode) of
group of machines 1 and 2 against that of machines 3, 4, 5, and 6; (2) the second
inter-area oscillation mode (second mode) of group of machines 3 and 4 against that
of machines 5 and 6; (3) a local oscillation mode (local mode) of machine 1 against
machine 2. Two other local oscillation modes are very well damped and hence are
not listed.
In the example power system, a multi-terminal UPFC (MUPFC) is installed at
node 7. The MUPFC is of one shunt and two series VSCs and its mathematical
model is given as:
387
G6
13
L 14
416MW
G1
160km
25km
7
MUPFC
160 km
14
15
16
10 km 25km
10
11
10km 25km
12
G5
G3
405MW
L9
2 L7
G2
G4
Fig. 7.29 A 6-machine power system installed with a MUPFC
Table 7.5 Oscillation modes in the example power system of Fig. 7.28
With the rst stabilizer
With the rst and second
stabilizers
With three stabilizers
First mode
Second mode
Local mode
0.115 j4.296
0.677 j3.987
0.827 j3.806
0.113 j3.338
0.059 j3.395
0.557 j3.588
0.230 j5.845
0.234 j5.825
0.234 j5.828
0.635 j3.792
0.610 j3.663
0.933 j5.623
dvdc 3mE
cos dE
dt
4Cdc
sin dE
3mB b
cos dB
4Cdc
iEd
3mB a
cos dB
4Cdc

iBd b
b
iBq b
iEq
sin dB
cos dB
6
vBtd a
0
xB a iBd a
2
6
4 mB a sin dB
vBtq a
iBq a
xB a
0
2
2
m
cos
dB
B
b

6
vBtd b
0
xB b iBd b
2
6
4
mB b sin dB
vBtq b
iBq b
xB b
0
2
2
3
mE cos dE vdc

6
7
vEtd
0 xE iEd
2
7
6
4 mE sin dE vdc 5
vEtq
iEq
xE
0
2
mB

iBd

a
iBq
sin dB
a vdc
7
7;
5
a vdc
b vdc
7
7;
5
v
b dc
a
a
388
where mE ; mB a ; mB b and dE ; dB a ; dB b are the modulation ratios and phase

angles of shunt and series VSCs, respectively. The MUPFC is to perform the
primary control functions to regulate the voltage at node 7 and power flow along
line 79 and line 714 with the following six PI controllers.
DC voltage regulator: dE dE0 Kdcp Ksdci vdcref vdc
AC voltage regulator: mE mE0 Kvp Ksvi Vref V7
Active and reactive power regulators for line 79:
dB
mB

KBapi
P79ref P79
a
a0
Bapp
s

KBaqi
Q79ref Q79
a
B a0
Baqp
s
dB
Active and reactive power regulators for line 714:

dB
mB

KBbpi
P714ref P714
b
b0
Bbpp
s

KBbqi
Q714ref Q714
b mB b0 KBbqp
s
dB
In order to improve the damping of those three oscillation modes, three stabilizers are to be added on three of six MUPFC normal control functions. A linearized
model of the system can be established by using the similar procedure presented in
Sect. 7.2. Based on the linearized model, controllability index to superimpose
damping function on the normal control functions of the MUPFC is calculated and
presented in Table 7.6. Maximal controllability predicts the best damping effect
(bold gures in Table 7.6). Hence, the AC voltage control function is selected to
add the rst stabilizer and the reactive power control function for line 79 is to add
the second stabilizer to damp the rst and second inter-area oscillation mode,
respectively. DC voltage control function is chosen to add the third stabilizer to
suppress the local oscillation mode. Locally available signals of active power are
used as the feedback signals and those three stabilizers are set in coordination using
the AFSA. Results of the coordinated design are given in Table 7.7.
Table 7.6 Controllability index of the MUPFC normal control functions
Control functions
1st mode
2nd mode
Local mode
AC voltage
Reactive power of line 79
Reactive power of line 714
DC voltage
Active power of line 79
Active power of line 714
62.06
9.86
9.04
72.59
0.89
0.88
6.93
19.52
17.85
5.47
0.98
0.99
9.93
0.86
0.91
16.07
0.18
0.18
389
Table 7.7 Results of coordinated design (T2 = T4 = 0.05, Tw = 5)

Parameters
The rst stabilizer
The second stabilizer
The third stabilizer
T1
T3
K
0.05
0.3
0.6
0.5
0.1
2.0
0.3
0.05
20.0
Success of coordinated design is conrmed by the results of computation of

oscillation modes presented in Table 7.5. From the third, fourth, and fth rows of
Table 7.5, it can be seen that those three stabilizers are set to improve the damping
of three oscillations of the example power system separately and the multi-mode
oscillations are damped effectively by their joint actions.

when only the rst stabilizer
is installed to damp the rst
inter-area oscillation mode.
a The rst mode inter-area
oscillation, b the second mode
inter-area oscillation, and
c local-mode oscillation
(a)
13
0.6
Without stabilizer
0.5
With the 1st stabilizers
0.4
0.3
0.2
0.1
0
-0.1
Time (second)
0
(b)
15
Without stabilizer
35
-0.01
10
-0.02
-0.03
-0.04
-0.05
-0.06
-0.07
-0.08
-0.09
Time (second)
0
(c)
0.24
0.22
0.2
0.18
0.16
0.14
0.12
0.1
0.08
0.06
0.04
12
10
15
Without stabilizer
Time (second)
0
10
15
390
Figures 7.30, 7.31, and 7.32 give the results of nonlinear simulation of the
example power system when it is subject to a three-phase to-earth fault at node 7 for
100 ms. They conrm the results of eigensolution in Table 7.5 that each stabilizer is
designed to damp the targeted oscillation mode and with three stabilizers
three-mode oscillations are suppressed successfully.

when only the rst and second
stabilizers are installed to
damp the rst and second
inter-area oscillation modes.
a The rst mode inter-area
oscillation, b the second mode
inter-area oscillation, and
c local-mode oscillation
(a)
0.6
13
0.5
Without stabilizer
With the 1st and 2nd stabilizers
0.4
0.3
0.2
0.1
0
-0.1
Time (second)
0
(b)
-0.01
-0.02
-0.03
-0.04
-0.05
-0.06
-0.07
-0.08
-0.09
35
10
15
Without stabilizer
Time (second)
0
(c)
0.24
0.22
0.2
0.18
0.16
0.14
0.12
0.1
0.08
0.06
0.04
12
10
15
Without stabilizer
Time (second)
0
10
15
References
Fig. 7.32 Simulation when
all three stabilizers are
installed. a The rst mode
inter-area oscillation, b the
second mode inter-area
oscillation, and c local-mode
oscillation
391
(a)
13
Without stabilizer
0.6
0.5
With all three stabilizers installed
0.4
0.3
0.2
0.1
0
-0.1
Time (second)
0
(b)
15
Without stabilizer
35
-0.01
10
-0.02
-0.03
-0.04
-0.05
-0.06
-0.07
-0.08
-0.09
Time (second)
0
10
15
(c)
0.24
0.22
0.2
0.18
0.16
0.14
0.12
0.1
0.08
0.06
0.04
12
Without stabilizer
Time (second)
0
10
15
References
1. CIGRE TF 30-01-08 Report (1999) Modelling of power electronics equipment (FACTS) in load
flow and stability programs
2. Wang HF (1999) Phillips-Heffron model of power systems installed with STATCOM and
applications. IEE Proc Part C 146(5):521527
3. Du W (2009) Power system small signal oscillation stability as affected by static synchronous
compensator (STATCOM) and energy storage system (ESS). PhD thesis, University of Bath,
UK
4. Wang HF (1999) Applications of modelling UPFC into multi-machine power systems. IEE Proc
Part C 3
5. Wang HF (2000) A unied model for the analysis of FACTS devices in damping power system
oscillations part III: unied power flow controller. IEEE Trans Power Del 3
392
6. Du W, Wang HF, Cao J, Xiao LY (2012) Application of the phase compensation method for the
design a DC/AC converter based stabilizer to damp power system inter-area oscillation. IEEE
Trans Power Syst 27(3):13021310
7. Du W, Wang HF, Cao J, Xiao LY (2012) Robustness of an energy storage system based
stabilizer to suppress inter-area oscillation in a multi-machine power system. IET Proc Gener
Transm Distrib 6(4):339351
8. Du W, Wu X, Wang HF, Dunn R (2010) Feasibility study to damp power system multi-mode
oscillations by using a single FACTS device. Int J Electr Power Energy Syst 32(6):645655
Index
A
AFSA, 381
AVR, 18
C
Closed-loop system, 32
Coordinated design, 204
Correlation ratio, 199
D
Damping control, 1, 81
Damping torque analysis (DTA), 3, 42, 81,
121, 266
Decentralized control, 209
E
Eigensolution, 4, 199
Eigenvalue, 4, 32
Electric torque, 18
Electromechanical oscillation, 38, 90
Electromechanical oscillation mode, 2, 189
ESS, 329
Excitation system, 18
Extended HeffronPhillips model, 81, 121
F
FACTS stabilizers, 8
Feedback signal, 245
Flexible AC transmission systems (FACTS), 5
Forward path, 90
G
General linearized model, 242, 336
H
HeffronPhillips model, 26, 189, 235, 329
I
Initial compensation, 88
Installing location, 97, 246
Inter-area line power oscillation, 349
L
Length of transmission line, 96
Linearized model, 17
Load condition, 92
Localized phase compensation, 349
Localized small-signal model, 350
M
Modal analysis (MA), 3, 199
Mode shape, 201
Modulation signal, 155
Multi-machine power system, 183, 235, 329
N
Non-negatively interactive design, 285
O
Optimum searching, 225
P
PAM control, 352
Participation factor, 200, 272
Phase compensation, 47, 139
Power system oscillations, 1
Power system stabilizer (PSS), 2, 17, 183, 245
PWM, 121
PWM control, 332
R
Residue index, 276
Robustness, 161

DOI 10.1007/978-1-4899-7696-3
393
394
S
Sensitivity index, 272
Single-machine innite-bus power system, 17,
81, 121
Small-signal angular stability, 1
SOFC, 367
SSSC, 133
SVC, 81, 235
T
TCPS, 97, 238
TCSC, 98, 238
Index
Transfer function, 19
U
UPFC, 143, 339
V
Voltage source converter (VSC), 6
VSC-based active power stabilizer, 130
VSC-based reactive power stabilizer, 130
VSC-based stabilizer, 121, 329

Analysis and Damping Control of Power System Low-Frequency Oscillations

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

Analysis and Damping Control of Power System Low-Frequency Oscillations

Загружено:

Авторское право:

Доступные форматы

Power Electronics and Power Systems

Power Electronics and Power Systems

More information about this series at http://www.springer.com/series/6403

Haifeng Wang Wenjuan Du

Analysis and Damping

Power system electromechanical low-frequency oscillations threaten the stable

2 A Single-Machine Innite-Bus Power System Installed

2.4.4 Equivalence Between the Damping Torque

4.3 Damping Function of a Unied Power Flow Controller (UPFC)

6.2 Analysis and Damping Control of Thyristor-Based FACTS

7.4 Damping of Multi-mode Oscillations by Multiple

Power System Low-frequency Oscillations

Power system low-frequency oscillations are the oscillations of active power

Springer Science+Business Media New York 2016

1.1 Power System Low-frequency Oscillations

Linearized Methods for the Analysis and Damping

Study of power system oscillations is in order to investigate the phenomena and to

1.3 FACTS and Grid-Connected ESS

FACTS and Grid-Connected ESS

Flexible AC transmission systems (FACTS) are the technologies emerged around

Fig. 1.1 Illustration on the

Three-phase diagram of GTO-based

Single-line diagram of GTO-based

Fig. 1.2 An SVS realized by a VSC circuit

1.3 FACTS and Grid-Connected ESS

Controllers to Damp Power System Oscillations

Through the effort of and experience accumulated by power engineers and

1.4 Controllers to Damp Power System Oscillations

 s is the voltage measurement at the location where the SVC

Occurrence of power system oscillations is essentially due to the response of

Design of Damping Controllers to Suppress Power

The prime specication of designing a damping controller (a PSS or a FACTS

1.5 Design of Damping Controllers to Suppress Power System Oscillations

In order to solve the problem of eigenvalue drifting caused by interactions

Organization of the Book

1.6 Organization of the Book

Table 1.1 Chapter arrangement of the book

A Single-Machine Innite-Bus Power

Linearized Model of a Single-Machine Innite-Bus

General Linearized Mathematical Model

Full Mathematical Model of a Synchronous Generator

Fundamental equations describing the dynamics of a synchronous generator are

2 A Single-Machine Innite-Bus Power System

of synchronous generator. Since there is no static coupling between any d-axis

where xd ; xq ; xf ; xD ; xQ are the self-reactance of winding d, q, f, D, and Q,

Excitation System and the Automatic Voltage Regulator

Voltage control in a power system is closely related to the regulation of reactive

2.1 Linearized Model of a Single-Machine

system and is implemented by solving some optimization problems. A synchronous

2 A Single-Machine Innite-Bus Power System

2.1 Linearized Model of a Single-Machine

rotational armature and

Fig. 2.5 Arrangement of a potential-source controlled-rectier excitation system

2 A Single-Machine Innite-Bus Power System

From Fig. 2.1, it can have

A Single-Machine Innite-Bus Power System

Fig. 2.6 A single-machine innite-bus power system

2.1 Linearized Model of a Single-Machine

Fig. 2.7 dq coordinate of

where vd Vb sin d; vq Vb cos d and the terminal voltage of generator is as

2 A Single-Machine Innite-Bus Power System

Linearized Model of Single-Machine Innite-Bus Power

Linearization of Eqs. (2.1) and (2.2) is as follows:

s is the voltage measurement at the location where the SVC

Equation (2.48) is the following polynomial equation if Ao is an M M matrix

xk t vk1 z1 0ek1 t vk2 z2 0ek2 t vkM zM 0ekM t

zi t w1i x1 t w2i x2 t wMi xM t