You are on page 1of 403

Power Electronics and Power Systems

Haifeng Wang
Wenjuan Du

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

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 Wenjuan Du
Beijing Beijing
China China

ISSN 2196-3185 ISSN 2196-3193 (electronic)


Power Electronics and Power Systems
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, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms 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 specific 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 Haifeng Wang


October 2015 Wenjuan Du

v
Contents

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........... 1
1.1 Power System Low-frequency Oscillations . . . . . . ........... 1
1.2 Linearized Methods for the Analysis and Damping
Control of Power System Oscillations . . . . . . . . . ........... 3
1.3 FACTS and Grid-Connected ESS . . . . . . . . . . . . ........... 5
1.4 Controllers to Damp Power System Oscillations . . ........... 7
1.5 Design of Damping Controllers to Suppress Power
System Oscillations . . . . . . . . . . . . . . . . . . . . . . ........... 10
1.6 Organization of the Book . . . . . . . . . . . . . . . . . . ........... 12
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........... 13
2 A Single-Machine Infinite-Bus Power System Installed
with a Power System Stabilizer. . . . . . . . . . . . . . . . . . . . ........ 17
2.1 Linearized Model of a Single-Machine Infinite-Bus
Power System Installed with a Power System Stabilizer . . . . . . . . 17
2.1.1 General Linearized Mathematical Model. . . . . . . . . . . . . . 17
2.1.2 Heffron–Phillips Model . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.2 Modal Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.2.1 Basis of Modal Analysis Theory . . . . . . . . . . . . . . . . . . . 32
2.2.2 Applications of Modal Analysis . . . . . . . . . . . . . . . . . . . 37
2.3 Damping Torque Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.3.1 Damping Torque and Synchronizing Torque . . . . . . . . . . . 42
2.3.2 Damping Torque Analysis and Design
of PSS by Phase Compensation. . . . . . . . . . . . ..... ... 47
2.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... ... 53
2.4.1 Linearized Mathematical Models of an Example Power
System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... ... 53
2.4.2 Modal Analysis of Example Power System. . . . ..... ... 59
2.4.3 Damping Torque Analysis of Example
Power System . . . . . . . . . . . . . . . . . . . . . . . . ........ 66

vii
viii Contents

2.4.4 Equivalence Between the Damping Torque


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

4.3 Damping Function of a Unified Power Flow Controller (UPFC)


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

6.2 Analysis and Damping Control of Thyristor-Based FACTS


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

7.4 Damping of Multi-mode Oscillations by Multiple


Stabilizers Attached to a Single UPFC . . . . . . . . . . . . . . . . . . . . 380
7.4.1 Coordinated Design of Multiple Stabilizers
by Artificial Fish Swarm Algorithm (AFSA) . . . . . . . . . . . 381
7.4.2 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 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 oscilla-
tions 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 sec-
onds 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.

© Springer Science+Business Media New York 2016 1


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
2 1 Introduction

Power system oscillation was first 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 trans-
mission 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 con-
ditions (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 oscil-
lation 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 classified,
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 3

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 oscil-
lation 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 infinite-bus power system to investigate the effect of excitation
control on power system small-signal angular stability. Reference [6] is the mile-
stone contribution by deMello and Concordia to the field. Proposal of the DTA is
based on the linearized Heffron–Phillips model of the single-machine infinite-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, synchro-
nizing 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.
4 1 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 [9–11],
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 infinite-bus and a multi-machine power system.
Eigensolution is a very specific 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 identification 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 difficulty 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 [13–15], 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 difficult 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 5

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:

V1 V2
P12 ¼ sin h ð1:1Þ
x

where h is the phase difference between the voltage at the sending and receiving end
of the transmission line, V  1 and V
 2 . A FACTS controller is to change in real time
the line impedance x, and phase and magnitude of line voltage V  1 and V 2 indi-
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 profile. 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
classified 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 V1 V2


operational principle of x
FACTS controllers

P12
6 1 Introduction

1. Thyristor-controlled static VAR compensator (SVC), which controls the mag-


nitude 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 modu-
lation (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 real-
ized 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 unified 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 Single-line diagram of GTO-based


voltage source converter voltage source converter SVS

Vc Vdc

transformer
Variable AC DC
voltage voltage
Power switching A DC m φ
circuit capacitor

Fig. 1.2 An SVS realized by a VSC circuit


1.3 FACTS and Grid-Connected ESS 7

Step-down transformer

Vc Vc
φ φ
ac/dc VSC
m m

Cdc ESS
nd
2 converter
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, stor-
age, and generation.
Figure 1.3 shows the configuration 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 configuration 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 supple-


mentary damping controller, the PSS (power system stabilizer), on the side of
excitation system of a generator is a simple, effective and economical method.
8 1 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 sufficient 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 imple-
mented, 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 field tests
confirmed 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
(fixed capacitor) type of SVC with a normal function of voltage control. A damping
controller (the SVC stabilizer) is superimposed on the SVC voltage control func-
tion. In Fig. 1.4, the variable inductive current Il is controlled by the thyristor firing
circuit. Hence, the variable equivalent reactance of the TCR-FC SVC is as follows
1.4 Controllers to Damp Power System Oscillations 9

Power system

Vs
Is

Il Ic
x svcl
α0
Vs π

Vsref - x svcc
voltage α firing circuit
controller +
+

π/2

ysvc SVC
stabilizer

Fig. 1.4 An SVC stabilizer superimposed on the SVC voltage control function

s
V s
V
jxsvc ¼  ¼   ð1:2Þ
Is ðIl þ Ic Þ

where and in Fig. 1.4, V  s is the voltage measurement at the location where the SVC

is installed, Vsref the reference signal of voltage control, jxsvc the equivalent reac-
tance 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
location close to V  sref by controlling the variable equivalent reactance of the SVC.
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
confirmed by case study of numerical calculation, computer simulation, and field
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 1 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 classified 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 oscilla-
tions, which have been demonstrated by simulation and laboratory experiment [25,
26].
In [27], a case of field 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 specification 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 sta-
bilizer is concerned, justification 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 justification 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
coefficients (IDSTCs) index [28], damping torque coefficient (DTC) index [29],
simple index [30], and damping index [31] derived by using the DTA. Sensitivity
calculation [32], the participation factors [13–15], 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 classified 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 firstly and its
parameters are set. Afterwards, stabilizer B is installed with its parameters being set.
The strategy of sequence setting fits 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 [9–11] 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 con-
sequence is that the sequential setting cannot complete the required setting of two
stabilizers, A and B, to move oscillation modes to their required positions simul-
taneously. This phenomenon of “eigenvalue drifting” is also referred to as the
problem of dynamic interactions between multiple stabilizers.
12 1 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 posi-
tions. Simultaneous tuning is often called the coordinated design of multiple sta-
bilizers. 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 classified into two groups. The first 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 opti-
mization. They usually convert the simultaneous tuning to the solution of an
optimization problem by establishing an objective function in the space of stabi-
lizers’ parameters. Then, an optimization algorithm is employed to find 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 them-
selves 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 oscil-
lations 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 Single-machine infinite-bus Multi-machine power
Type of power systems power systems systems
The PSSs Chapter 2 Chapter 5
The thyristor-based FACTS Chapter 3 Chapter 6
stabilizers
The VSC-based FACTS Chapter 4 Chapter 7
stabilizers

analysis (DTA) and the modal analysis (MA). The basic concepts and theory of the
DTA and the MA are presented firstly for a simpler case of power systems, a
single-machine infinite-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 classified 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 classification of damping controllers (stabilizers) and their applications in the
simple single-machine infinite-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 Definitions (2004) Definition and
classification of power system stability. IEEE Trans Power Syst 19(2):1387–1401
2. Schleif FR, White JH (1966) Damping for the northwest-southwest tieline oscillations—an
analogue study. IEEE Trans Power Appar Syst 85(12):1239–1247
3. Gibson CM (1988) Application of power system stabilizers on the Anglo-Scottish
interconnection—programme of system proving tests and operational experience. IEE Proc
Part C 135(3):255–260
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):92–98
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):316–329
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:
692–697
8. deMello FP, Laskowski TF (1975) Concepts of power system dynamic stability. IEEE Trans
Power Appar Syst 94(3):827–833
9. Larsen EV, Swann DA (1981) Applying power system stabilizers part I: general concepts.
IEEE Trans Power Appar Syst 100(6):3017–3024
14 1 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):3025–3033
11. Larsen EV, Swann DA (1981) Applying power system stabilizers part III: practical
considerations. IEEE Trans Power Appar Syst 100(6):3034–3046
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):3117–3125
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):3126–3134
15. Sancha JL, Perez-Arriaga IJ (1988) Selective modal analysis of electric power system
oscillatory instability. IEEE Trans Power Syst 3(2):429–438
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):948–956
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):569–574
18. Hingorani NG (1988, Aug) High power electronics and flexible ac transmission systems. IEEE
Power Eng Rev 3–4
19. Gyugyi L (1992) Unified power-flow control; concept for flexible AC transmission systems.
IEE Proc Part C 139(4):323–331
20. Smith OJM (1969) Power system transient control by capacitor switching. IEEE Trans Power
Appar Syst 88(1):28–35
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):427–433
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):990–996
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 1524–1529
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):
145–151
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):1920–1925
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):359–364
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):155–159
32. Abdalla OH, Hassan SA, Tweig NT (1984) Co-ordinated stabilisation of a multimachine
power system. IEEE Trans Power Appar Syst 103(3):483–494
33. Okamoto H, Kurita A, Sekine Y (1995) A method for identification of effective locations of
variable impedance apparatus on enhancement of steady-state stability in large-scale power
systems. IEEE Trans Power Syst 10(3):1401–1407
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
systems. IEEE Trans Power Syst 5(4):1455–1469
35. Yang X, Feliachi A (1994) Stabilization of inter-area oscillation modes through excitation
systems. IEEE Trans Power Syst 9(1):494–502
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):3879–3887
Chapter 2
A Single-Machine Infinite-Bus Power
System Installed with a Power System
Stabilizer

2.1 Linearized Model of a Single-Machine Infinite-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 Park’s voltage equations. They are given based on a coordinate
system consisting of a d-axis (direct axis) fixed on the field winding axis of syn-
chronous generator and a q-axis (quadrature axis). After Park’s 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. Park’s voltage equations of those five windings have the simplest form
as follows:

w_ d ¼ x0 ðvtd þ ra id þ xwq Þ
w_ q ¼ x0 ðvtq þ ra iq  xwd Þ
w_ ¼ x0 ðvf  rf if Þ
f ð2:1Þ
w_ D ¼ x0 rD iD
w_ Q ¼ x0 rQ iQ

where wd ; wq ; wf ; wD ; wQ , id ; iq ; if ; iD ; iQ , vtd ; vtq ; vf , and ra ; rf ; rD ; rQ are the mag-


netic 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.)

© Springer Science+Business Media New York 2016 17


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_2
18 2 A Single-Machine Infinite-Bus Power System …

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 3 2 32 3
wd xd xad xad id
6 7 6 76 7
4 wf 5 ¼ 4 xad xf xad 54 if 5
wD xad xad xD iD ð2:2Þ
" #   
wq xq xaq iq
¼
wQ xaq xQ iQ

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 ¼ x0 ðx  1Þ
: 1 ð2:3Þ
x ¼ ½Tm  Tt  Dðx  1Þ
M

where M is the inertia of the rotor, D the damping coefficient 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 profile 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 trans-
mission 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 refer-
ence values. The highest level of voltage control uses global information of power
2.1 Linearized Model of a Single-Machine … 19

Fig. 2.1 Arrangement of an if


AVR

vf0 vf

Vt
-
Vtref
vf' TE(s)
+
+ 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 syn-
chronous generator.
In Fig. 2.1, TE(s) denotes the transfer function of the AVR. The excitation
provided to the field 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 generator’s terminal. On the other hand, the refer-
ence 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 profile of the
system is kept within required constraints and the transmission losses are mini-
mized. 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 generator’s terminals and were implemented manually. Gradually,
fast-acting AVRs were installed in many generators not only providing a satis-
factory 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 2 A Single-Machine Infinite-Bus Power System …

synchronous generator
DC motor

slip ring

field winding of - Vt
dc motor 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 classified into three major types.
1. DC excitation systems
A DC excitation system uses a DC generator as the source of excitation power to
provide field current and is connected to field 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 exci-
tation power to provide field current. The AC excitation current is rectified 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 rectifier excitation system, the armature
windings of the AC exciter and the diode rectifiers rotate with the synchronous
generator field. 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 field of generator through slip rings after being
rectified. There are three major types of static excitation systems. They are
potential-source controlled-rectifier systems, compound-source rectifier systems,
and compound-controlled rectified excitation systems. Figure 2.5 shows the
arrangement of a potential-source controlled-rectifier excitation system.
2.1 Linearized Model of a Single-Machine … 21

alternator stationary diode synchronous generator

slip ring
Vt
field winding of -
alternator
Vtref
AVR
+

Fig. 2.3 Arrangement of a stationary rectifier excitation system

rotational armature and


alternator diode rectifier synchronous generator

Vt
field winding of − Vtref
alternator
AVR
+

Fig. 2.4 Arrangement of a rotating rectifier excitation system

controlled rectifier

Vt

Vtref
AVR
+

Fig. 2.5 Arrangement of a potential-source controlled-rectifier excitation system


22 2 A Single-Machine Infinite-Bus Power System …

From Fig. 2.1, it can have

vf ¼ vf0 þ TEðsÞðVtref  Vt þ upss Þ ð2:4Þ

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, respec-
tively. 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,

KA
TEðsÞ ¼ ð2:5Þ
1 þ sTA

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 ¼ vf0 þ v0f
1 KA   ð2:6Þ
v_ 0f ¼  v0f þ Vtref  Vt þ upss
TA TA

2.1.1.3 A Single-Machine Infinite-Bus Power System

Figure 2.6 shows the configuration 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 “infinite”. Hence, busbar b is called the
“infinite busbar”, and the part of the power system on the left-hand side of busbar b
is a “single-machine infinite-bus” power system. The single-machine infinite-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 infinite-bus power system


2.1 Linearized Model of a Single-Machine … 23

Fig. 2.7 d–q coordinate of q


generator Vt

jxt I t

vq Vb

It
iq δ
d
id vd

For the single-machine infinite-bus power system shown in Fig. 2.6, it can have

 t ¼ jxtIt þ V
V b ð2:7Þ

In the d–q coordinate of the generator shown in Fig. 2.7, Eq. (2.7) can be written
as follows:

vtd þ jvtq ¼ jxt ðid þ jiq Þ þ vd þ jvq ð2:8Þ

where vtd ; vtq and id ; iq vd ; vq are the d and q components of terminal voltage of
generator, V t , line current, It , and voltage at the infinite busbar, V
 b , respectively.
Comparing the real and imaginary part on the both sides of Eq. (2.8), it can have

vtd ¼ xt iq þ vd
ð2:9Þ
vtq ¼ xt id þ vq

where vd ¼ Vb sin d; vq ¼ Vb cos d and the terminal voltage of generator is as


follows:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Vt ¼ v2td þ v2tq ð2:10Þ

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 ¼ Pm and Tt ¼ Pt . While Pt is equal to the
electric power received at the infinite busbar, that is

Pt ¼ vdt id þ vqt iq ¼ vd id þ vq iq ð2:11Þ


24 2 A Single-Machine Infinite-Bus Power System …

Equations (2.1)–(2.3), (2.6) and (2.9)–(2.11) are the complete dynamic model of
the single-machine infinite-bus power system shown in Fig. 2.6 where Vb and Pm
are constant.

2.1.1.4 Linearized Model of Single-Machine Infinite-Bus Power


System

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

Dw_ d ¼ x0 ðDvtd þ ra Did þ x0 Dwq þ wq0 DxÞ


Dw_ q ¼ x0 ðDvtq þ ra Diq  x0 Dwd  wd0 DxÞ
Dw_ ¼ x0 ðDvf  rf Dif Þ
f ð2:12Þ
Dw_ D ¼ x0 rD DiD
Dw_ Q ¼ x0 rQ DiQ
2 3 2 31 2 3
Did xd xad xad Dwd
6 7 6 7 6 7
4 Dif 5 ¼ 4 xad xf xad 5 4 Dwf 5
DiD xad xad xD DwD ð2:13Þ
   1 " #
Diq xq xaq Dwq
¼
DiQ xaq xQ DwQ

where prefix, 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 nota-
tion will be used throughout this book.
Linearization of Eqs. (2.10) and (2.11) is as follows:
vtd0 vtq0
DVt ¼ Dvtd þ Dvtq ð2:14Þ
Vt0 Vt0

DPt ¼ vdt0 Did þ vqt0 Diq þ id0 Dvdt þ iq0 Dvqt ð2:15Þ

By using Eqs. (2.14) and (2.15), linearization of Eqs. (2.3) and (2.6) can be
obtained to be

Dd_ ¼ x0 Dx
1
Dx_ ¼  ðDPt þ DDxÞ ð2:16Þ
M
1
¼  ðvdt0 Did þ vqt0 Diq þ id0 Dvdt þ iq0 Dvqt þ DDxÞ
M
2.1 Linearized Model of a Single-Machine … 25

Dvf ¼ Dv0f
 
1 KA vtd0 vtq0 ð2:17Þ
Dv_ 0f 0
¼  Dvf þ  Dvtd  Dvtq þ Dupss
TA TA Vt0 Vt0

Arranging Eqs. (2.12), (2.13), (2.16), and (2.17) in matrix form with all lin-
earized current variables be cancelled, it can have

sDXgdq ¼ Agdq DXgdq þ Bgdq DVdq þ bpss Dupss


ð2:18Þ
DIdq ¼ Cgdq DXgdq

where

T
DXgdq ¼ Dd Dx Dvf Dwd Dwq Dwf DwD DwQ ;
T T
DVdq ¼ ½ Dvtd Dvtq  ; DIdq ¼ ½ Did Diq 

For the single-machine infinite-bus power system, the network voltage equation
is Eq. (2.9). Its linearization is as follows:

Dvtd ¼ xt Diq þ Dvd ¼ xt Diq þ Vb cos d0 Dd


ð2:19Þ
Dvtq ¼ xt Did þ Dvq ¼ xt Did  Vb sin d0 Dd

In matrix form, the above equation can be written as follows:

DVdq ¼ Fdq1 DIdq þ Fdq2 DXgdq ð2:20Þ

where
   
0 xt Vb cos d0 0
Fdq1 ¼ ; Fdq2 ¼
xt 0 Vb sin d0 0

Substituting Eq. (2.20) into (2.18), state-equation model of the single-machine


infinite-bus power system is obtained to be

sDXgdq ¼ Agcdq DXgdq þ bpss Dupss ð2:21Þ

where Agcdq ¼ Agdq þ Bgdq Fdq1 Cgdq þ Bgdq Fdq2 :


26 2 A Single-Machine Infinite-Bus Power System …

2.1.2 Heffron–Phillips Model

2.1.2.1 Simplification

For the study of power system oscillations, full mathematical model of synchronous
generator of Eqs. (2.1)–(2.2) can be simplified based on the following
considerations:
1. Effect of damper windings is not considered or directly included in the damping
coefficient D in the rotor motion equation in Eq. (2.3). Thus, Eq. (2.1) is sim-
plified to be
w_ ¼ x0 ðvtd þ ra id þ xw Þ
d q

w_ q ¼ x0 ðvtq þ ra iq  xwd Þ ð2:22Þ


w_ f ¼ x0 ðvf  rf if Þ

2. Effect of fast transient and the resistance of d and q armature windings are
neglected. Equation (2.22) is further simplified to be

0 ¼ vtd þ xwq
0 ¼ vtq  xwd ð2:23Þ
w_ f ¼ x0 ðvf  rf if Þ

3. In small-signal power oscillations, variation of rotor speed is very small, x  1.


Hence, the first two equations in Eq. (2.23) become

vtd ¼ wq
ð2:24Þ
vtq ¼ wd

To transform the third equation in Eq. (2.23) into a different form, it is defined
that
xad xad vf
E0q ¼ w ; Eq ¼ xad if ; Efd ¼ ð2:25Þ
xf f rf

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
2.1 Linearized Model of a Single-Machine … 27

0
T0d0 E_ q ¼ Efd  Eq ð2:26Þ

where T0d0 ¼ xx0 rf f , which is the time constant of the field winding.
Equation (2.26) together with Eq. (2.3) forms the simplified third-order model of
synchronous generator. Equation (2.2) becomes
    
wd xd xad id
¼
wf xad xf if ð2:27Þ
wq ¼ xq iq

From Eqs. (2.24), (2.25), and (2.27), it can have

vtd ¼ wq ¼ xq iq
ð2:28Þ
vtq ¼ wd ¼ xad if  xd id ¼ Eq  xd id

From Eqs. (2.25) and (2.27), it can be obtained that

xad xad x2  
E0q ¼ wf ¼ ðxf if  xad id Þ ¼ Eq  ad id ¼ Eq  xd  x0d id ð2:29Þ
xf xf xf

x2
where x0d ¼ xd  xadf , which is called the transient d-axis reactance. Thus, Eq. (2.26)
becomes
0  
T0d0 E_ q ¼ Efd  E0q  xd  x0d id ð2:30Þ

2.1.2.2 A Simplified Model of Single-Machine Infinite-Bus Power


System

For the single-machine infinite-bus power system shown in Fig. 2.6, from Eqs. (2.9)
and (2.28), it can have

vtd ¼ vd  xt iq ¼ xq iq
ð2:31Þ
vtq ¼ vq þ xt id ¼ Eq  xd id ¼ E0q  x0d id

Thus,

vd ¼ ðxt þ xq Þiq ¼ xqR iq


ð2:32Þ
vq ¼ Eq  ðxd þ xt Þid ¼ E0q  ðx0d þ xt Þid ¼ Eq  xdR id ¼ E0q  x0dR id
28 2 A Single-Machine Infinite-Bus Power System …

Fig. 2.8 Equivalent circuit iq id


model of single-machine + +
infinite-bus power system

x dΣ (x dΣ ')
x qΣ vd vq

Eq (Eq ')

d winding
- -
q winding

Fig. 2.9 Phasor diagram of q

single-machine infinite-bus EQ
power system

j(x q − x 'd )I t

Eq '
jx 'd I t

jx t I t
Vt b
vq Vb

δ
iq It
d

id vd

The single-machine infinite-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 d–q
coordinate.
From Eq. (2.32) or Fig. 2.8, it can have
E0q  Vb cos d
id ¼
x0dR
ð2:33Þ
Vb sin d
iq ¼
xqR

By substituting Eqs. (2.9) and (2.33) into Eq. (2.11), the electric power supplied
by the generator can be expressed as follows:
2.1 Linearized Model of a Single-Machine … 29

0
Vb sin d Eq  Vb cos d
Pt = Vb cos d þ Vb sin d
xqR x0dR
0   ð2:34Þ
Eq Vb V2 xq  x0d
¼ 0 sin d  b 0 sin 2d
xdR 2 xdR xqR

From Eqs. (2.29) and (2.33), it can be obtained that

    E0q  Vb cos d
Eq ¼ E0q þ xd  x0d id ¼ E0q þ xd  x0d
x0dR
0  0
 ð2:35Þ
Eq xdR xd  xd Vb cos d
¼ 0 
xdR x0dR

where xdR ¼ xd þ xt . From Eqs. (2.31) and (2.33), it can have

Vb sin d xq Vb sin d
vtd = Vb sin d  xt iq ¼ Vb sin d  xt ¼ ;
xqR xqR
ð2:36Þ
E0q  Vb cos d xt E0q Vb x0d cos d
vtq = Vb cos d þ xt id ¼ Vb cos d þ xt ¼ þ
x0dR x0dR x0dR

Hence, the simplified model of single-machine infinite-bus power system is as


follows:
:
d ¼ x0 ðx  1Þ
: 1
x ¼ ½Pm  Pt  Dðx  1Þ
M
0 1   ð2:37Þ
E_ q ¼ 0 Eq þ Efd
Tdo
0 1 KA  
E_ fd ¼  E0fd þ Vtref  Vt þ upss
TA TA

where
 
E0q Vb V2b xq  x0d
Pt ¼ 0 sin d  sin 2d
xdR 2 x0dR xqR
 
E0q xdR xd  x0d Vb cos d
Eq ¼ 0 
xdR x0dR ð2:38Þ
Efd ¼ Efd0 þ E0fd
xq Vb sin d xt E0q Vb x0d cos d qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
vtd ¼ ; vtq ¼ 0 þ 0 ; Vt ¼ v2td þ v2tq
xqR xdR xdR
30 2 A Single-Machine Infinite-Bus Power System …

The model is a group of 4 first-order differential equations plus 6 algebraic


equations.

2.1.2.3 Heffron–Phillips Model [1–3]

By linearizing Eqs. (2.37) and (2.38) at an operating point of power system, where
Vt ¼ Vt0 ; Vtd ¼ Vtd0 ; Vtq ¼ Vtq0 ; d ¼ d0 ; x0 ¼ 1; E0q ¼ E0q0 ; Efd ¼ Efd0 , it can have

Dd_ ¼ xo Dx
1
Dx_ ¼ ðDPt  DDxÞ
M
0 1 ð2:39Þ
DE_ q ¼ 0 ðDEq þ DE0fd Þ
Tdo
0 1 KA
DE_ fd ¼  DE0fd  ðDVt  Dupss Þ
TA TA

DPt ¼ K1 Dd þ K2 DE0q
DEq ¼ K3 DE0q þ K4 Dd ð2:40Þ
DVt ¼ K5 Dd þ K6 DE0q

where

E0q0 Vb x0dR V 2 ð xq  x0 Þ
K1 ¼ cos d0  bx0 xqR d cos 2d0
dR

K2 ¼ xV0 b sin d0
dR
K3 ¼ xxdR
0
dR
ðxd x0d ÞVb sin d0
K4 ¼ x0dR
X q Vb cos d0 V V x0 sin d0
K5 ¼ VVt0
td0
xqR  Vtq0 b0 xd0
t0 dR
V
K6 ¼ Vtq0 xx0 t
t0 dR

Substituting Eq. (2.40) into Eq. (2.39), it can be obtained that

Dd_ ¼ xo Dx
1
Dx_ ¼ ðK1 Dd  K2 DE0q  DDxÞ
M
_ 0 1 ð2:41Þ
DEq ¼ 0 ðK3 DE0q  K4 Dd þ DE0fd Þ
Tdo
0 1 KA
DE_ fd ¼  DE0fd  ðK5 Dd þ K6 DE0q  Dupss Þ
TA TA
2.1 Linearized Model of a Single-Machine … 31

K1

ΔPt Δω ω0 Δδ
_ 1
Ms + D s

K4 K5
K2
_ _
ΔEq’ 1 + KA + Δupss
Td0's+K3 1+sTA _

K6

Fig. 2.10 Heffron–Phillips model of a single-machine infinite-bus power system

Equation (2.41) is the so-called Heffron–Phillips model of single-machine


infinite-bus power system, which is shown in Fig. 2.10.
The Heffron–Phillips model can be written in the form of state-space repre-
sentation of Eq. (2.21) where
2 3
2 3 0 xo 0 0
Dd 6 7
6 Dx 7 6 K 1
MD K 2
0 7
6 7 6 M M 7
DXgdq ¼6 7; Agcdq ¼6
6  K0 4 1 7;
7
4 DE0q 5 6 0  K0 3
4 Tdo Tdo T0do 7
5
DE0fd  TA K5
K 0  TA K6
K  T1 ð2:42Þ
2 3 A A A

0
6 7
6 0 7
bpss ¼6
6 0 7
7
4 5
KA
TA
32 2 A Single-Machine Infinite-Bus Power System …

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 ¼ Ao X þ bo u
y ¼ cTo X ð2:43Þ
u ¼ HðsÞy

where Ao , bo and cTo is the state matrix, control vector, and output vector of
open-loop system, respectively, and HðsÞ is the transfer function of feedback
controller. Transfer function of open-loop system is as follow:
y
GðsÞ ¼ ¼ cTo ðsI  Ao Þ1 bo ð2:44Þ
u

The system is shown in Fig. 2.11. Transfer function of closed-loop system is as


follows:

y GðsÞ
TðsÞ ¼ ¼ ð2:45Þ
w 1  GðsÞHðsÞ

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 satisfies the following equation

Ao v ¼ kv ð2:46Þ

with a non-trivial solution ðv 6¼ 0Þ.

Fig. 2.11 Block diagram of a closed-loop control system


2.2 Modal Analysis 33

Obviously, Eq. (2.46) can be written in the following form

ðAo  kIÞv ¼ 0 ð2:47Þ

where I is an unity matrix. In order for Eq. (2.47) to have the non-trivial solution, it
should have

jAo  kIj ¼ 0 ð2:48Þ

Equation (2.48) is the following polynomial equation if Ao is an M  M matrix

ð1ÞM kM þ aM1 kM1 þ    þ a1 k þ a0 ¼ 0 ð2:49Þ

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 satisfies the
equation

Avi ¼ ki vi ; i ¼ 1; 2; . . . ; M ð2:50Þ

vi is called the right eigenvector of matrix A associated with ki . Equation (2.50)


can be arranged as follows:
2 3
k1 0 0 0
60 k2 0 0 7
6 7
A½ v1 v2 . . . vM  ¼ ½ v1 v2 . . . vM 6 .. 7 ð2:51Þ
40 0 . 0 5
0 0 0 kM

That is

V1 AV ¼ K ð2:52Þ

where
2 3
k1 0 0 0
60 k2 0 0 7
6 7
V ¼ ½ v1 v2 ... vM ; K ¼ 6 .. 7
40 0 . 0 5
0 0 0 kM
34 2 A Single-Machine Infinite-Bus Power System …

Denote
2 3
wT1
6 wT 7
6 2 7
V 1
¼6
6 ..
7 ¼ ½ w1
7 w2 . . . w M  T ¼ WT ð2:53Þ
4 . 5
wTM

From Eqs. (2.52) and (2.53), obviously it can have

wTi A ¼ wTi ki ; i ¼ 1; 2; . . .; M ð2:54Þ

Hence, wTi is called the left eigenvector corresponding to eigenvalue ki .


If a new state variable vector z is introduced and defined to be

X ¼ VZ ð2:55Þ

From Eqs. (2.43), (2.52), and (2.53), it can be obtained that

sZ ¼ KZ þ WT bo u
ð2:56Þ
y ¼ cTo VZ

That is

szi ¼ ki zi ; þ wTi bo u i ¼ 1; 2; . . .; M
X
M ð2:57Þ
y ¼ cTo vi z i
i¼1

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.

1 z1
w1T b 0 c0 T v1
s − λ1

1 z2
w 2Tb0 c0 T v 2
s − λ2
u y
+

1 zM
w M Tb0 c0 T v M
s − λM

Fig. 2.12 Block diagram of modal decomposition representation of state-space model of


open-loop system
2.2 Modal Analysis 35

2.2.1.2 Stability of Open-Loop System and Closed-Loop System

Considering the open-loop system when u ¼ 0, the state-space representation of


Eq. (2.57) is as follows:

szi ¼ ki zi ð2:58Þ

Solution of Eq. (2.58) is as follows:

zi ðtÞ ¼ zi ð0Þeki t ; i ¼ 1; 2; . . .; M ð2:59Þ

where zi ð0Þ is the initial value of state variable zi ðtÞ; i ¼ 1; 2; . . .; M. From


Eqs. (2.55) and (2.59), it can have
2 3
z1 ð0Þek1 t
6 z2 ð0Þek2 t 7
6 7
X ¼ V6 .. 7 ð2:60Þ
4 . 5
zn ð0ÞekM t

Hence, time response of the kth state variable of the system,


xk ðtÞ; i ¼ 1; 2; . . .; M, is as follows:

X
M
xk ðtÞ ¼ vk1 z1 ð0Þek1 t þ vk2 z2 ð0Þek2 t þ    þ vkM zM ð0ÞekM t ¼ vki zi ð0Þeki t
i¼1
ð2:61Þ

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
conjugate complex number, i.e. k i;i þ 1 ¼ n  jxi , the corresponding component in
i
the time response of the kth state variable of the system should be

vki zi ð0Þeki t ¼ vki zi ð0Þeðni þ jxi Þt ¼ vki zi ð0Þeni t ½cos xi t þ j sin xi t
 ð2:62Þ

The component is oscillatory in respect of time. The oscillation angular fre-


quency is xi . The decaying and increasing of the oscillation are determined by the
36 2 A Single-Machine Infinite-Bus Power System …

real part of the mode ni . The pair of conjugate eigenvalues of state matrix Ao are
often called the oscillation mode of the system.
The oscillation frequency f i (Hz) and damping fi associated with k i;i þ 1 ¼
ni  jxi are normally defined as follows:

xi ni
fi ¼ ; f ¼  qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð2:63Þ
2p i n þ x2
2
i i

From Eq. (2.59), it can be seen that zi ðtÞ; i ¼ 1; 2; . . .; M is related only with the
ith mode of the system ki . Hence, zi ðtÞ; i ¼ 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 ðtÞ. 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 ¼ wTi bo cTo vi ð2:64Þ

From Eqs. (2.53) and (2.55), it can have

Z ¼ V1 X ¼ WT X ð2:65Þ

or

zi ðtÞ ¼ w1i x1 ðtÞ þ w2i x2 ðtÞ þ    þ wMi xM ðtÞ ð2:66Þ

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 ðtÞ on the ith
state variable zi ðtÞ, 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 HðsÞ be

sXf ¼ Af Xf þ bf y
ð2:67Þ
u ¼ cTf Xf
2.2 Modal Analysis 37

That is HðsÞ ¼ cTf ðsI  Af Þ1 bf . Thus, from Eqs. (2.43) and (2.67), the
state-space representation of closed-loop system can be obtained to be
      
sX Ao bo cTf X X
¼ ¼ Ac ð2:68Þ
sXf bf cTo Af Xf Xf

where Ac is the state matrix of closed-loop system. Obviously, based on the dis-
cussion 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 ¼ Ao þ HðsÞbo cTo ð2:69Þ

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 ðaÞ @Hðki ; aÞ T T @Hðki ; aÞ


¼ wTi vi ¼ wi bo co vi ¼ Ri ð2:70Þ
@a @a @a @a

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 infinite-bus power system expressed


by
 the Heffron–Phillips
 model shown in Fig. 2.10 without the PSS installed
Dupss ¼ 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 0 xo  Dd  0
¼ þ DE0q
sDx K 1
MD Dx KM
2
M
 
Dd ð2:71Þ
Dd ¼ ½ 1 0
Dx
DE0q ¼ Fdelta ðsÞDd
38 2 A Single-Machine Infinite-Bus Power System …

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 ðsÞ. Obviously, the state-space realization of Fdelta ðsÞ is as follows:
2 3 2 3
   K0 3 1  0   K0 4
sDE0q 6 Tdo 0
Tdo 7 DE Tdo 5
¼4 5
q
þ4 Dd
sDE0fd K
 TA K6
T 1 DE0
 TA K5
K
A A
fd
A ð2:72Þ
 0 
DEq
DE0q ¼ ½ 1 0 
DE0fd

According to Eq. (2.48), the modes of open-loop system can be found by solving
the following characteristic equation
     
0 xo 1 0 k xo D K1
D  k 0 1 ¼  K1 D  k ¼ k þ M k þ M xo ¼ 0
2
 K1 M M
M M
ð2:73Þ

Solution of Eq. (2.73) gives the oscillation mode of the single-machine


infinite-bus power system when DE0q ¼ 0, that is the case when the generator is
modelled as a constant voltage source. The oscillation mode is as follows:
2 ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
s 3

1;2 14 D D 2 4x0 K1 5
k ¼    ¼ no  jxNOF ð2:74Þ
2 M M M

The oscillation mode is related to the rotor motion of generator, i.e. state vari-
ables Dd and Dx. It is often called the electromechanical oscillation mode of the
power system.
From Eq. (2.71), it can have
      
sDd 0 xo Dd Dd
¼ K K D ¼ Ac ð2:75Þ
sDx  M  M2 Fdelta ðsÞ
1
M Dx Dx

where Ac is the state matrix of closed-loop system.


2.2 Modal Analysis 39

From Eq. (2.72), it can be obtained that


2 31 23
s þ K0 3  1
0  K0 4
6 Tdo T 7 4 Tdo 5
Fdelta ðsÞ ¼ ½ 1 0 4 5
do

KA K6
TA sþ T1
 TA K5
K
A A
2 32 3
s þ T1 1
 K0 4
1 T0do Tdo 5
4 54
A
¼  ½1 0 K K
s þ K0 3 s þ T1 þ KTA K6 10  TA 6 s þ K0 3  TA K5
K
Tdo A A Tdo A Tdo A

2 3
T0do TA h i  K0 4
s þ T1 1 4 Tdo 5
¼ T0do
ðsT0do þ K3 ÞðTA s þ 1Þ þ KA K6 A
 TA K5
K
A

ðTA s þ 1ÞK4 þ KA K5
¼
ðsT0doþ K3 ÞðTA s þ 1Þ þ KA K6
ð2:76Þ

Hence, according to Eq. (2.70), it can have

ðTA ki þ 1ÞK4 þ KA K5


 @  0
@ ki ð i do þ K3 ÞðTA ki þ 1Þ þ KA K6
k T
¼ Ri ð2:77Þ
@a @a

where a ¼ 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 infinite-bus power system with
the PSS installed is Eq. (2.21) which can be rearranged as follows:
2 3 2 32 3 2 3
sDd 0 x0 0 Dd 0
6 7 6 a21 a22 aT238 76 7 6 7
4 sDx 5 ¼ 4 54 Dx 5 þ 4 0 5Dupss
M M M
sDx38 a138 a238 A33 Dx38 bpss3
2 3
Dd ð2:78Þ
6 7
y ¼ Dx ¼ ½ 0 1 0 4 Dx 5
Dx38
Dupss ¼ Tpss ðsÞDx
40 2 A Single-Machine Infinite-Bus Power System …

a 21

1 Δω ω0 Δδ
- Ms + a 22 s
ΔTe

a 23-8 T a 23-8 a13-8

ΔX 3-8

( sI - A 33 ) −1 + bpss-3 Δu pss

Fig. 2.13 General linearized model of single-machine infinite-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
2 3

vi1
wTi ¼ wi1 wi2 wTi3 ; vi ¼ 4 vi2 5 ð2:79Þ
vi3

That is
2 3

0 x0 0

wi1 wi2 wTi3 4 a21 a22 aT238 i wi1


5¼k wi2 wTi3 ð2:80Þ
M M M
a138 a238 A33

It can have
a21 i w
 i2
w  Ti3 a138 ¼ k
þw  i1
M
a22 i w
x0 w  i2
 i1  w  Ti3 a238 ¼ k
þw  i2 ð2:81Þ
M
aT i w
 i2 238 þ w
w  Ti3 A33 ¼ k  Ti3
M
2.2 Modal Analysis 41

Hence,

aT238 
 Ti3 ¼ w
w  i2 ðki I  A33 Þ1 ð2:82Þ
M

From Eqs. (2.64), (2.78), and (2.82), the residue can be obtained to be
2 3 2 3
0 vi1

6 7 6 7
Ri ¼ wi1 wi2  Ti3 4
w 0 5½ 0 1 0 4 vi2 5
bpss3 vi3
ð2:83Þ
¼ wTi3 bpss3 vi2
aT i I  A33 Þ1 bpss3
¼ w  i2 vi2 238 ðk
M

2.2.2.3 Design of PSS by Pole Assignment

If the feedback signal and transfer function of the PSS to be designed is y and
TPSS ðsÞ respectively, Eq. (2.78) can be written more generally as follows:

sX ¼ AX þ bDupss
y ¼ cT X ð2:84Þ
Dupss ¼ Tpss ðsÞy

Transfer function of open-loop system is as follows:

GðsÞ ¼ cT ðsI  AÞ1 b ð2:85Þ

Characteristic equation of closed-loop control system is as follows:

1 þ GðsÞTpss ðsÞ ¼ 0 ð2:86Þ

If design of the PSS is to assign the electromechanical oscillation mode of the


single-machine infinite-bus power system to a target position, k c ¼ n  jxc ,
c

kc must be the solution of the characteristic equation of closed-loop control system
of Eq. (2.86). Hence, it should have

c ÞTpss ðk
1 þ Gðk c Þ ¼ 0 ð2:87Þ
42 2 A Single-Machine Infinite-Bus Power System …

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
lead–lag block,

ð1 þ saTÞ2
Tpss ðsÞ ¼ Kpss ð2:88Þ
ð1 þ sTÞ2

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 Heffron–Phillips Model

The damping torque analysis (DTA) was firstly introduced on the basis of the
Heffron–Phillips model for a single-machine infinite-bus power system to examine
the effect of excitation control, such as the AVR, on power system small-signal
stability [1–3]. 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 infinite-bus power system.
In the Heffron–Phillips 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 field 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 Heffron–Phillips model is obviously an
electric torque. Hence, from Fig. 2.14, it can have

K1

Electromechanical oscillation loop

Δω ω0 Δδ
_ 1
Ms + D s

ΔTe

Fig. 2.14 Electromechanical oscillation loop of generator


2.3 Damping Torque Analysis 43

D x0 K1 x0
s2 Dd þ sDd þ Dd þ DTe ¼ 0 ð2:89Þ
M M M

If firstly the contribution from the lower part of Heffron–Phillips 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

D x0 K1
s2 Dd þ sDd þ Dd ¼ 0 ð2:90Þ
M M

Equation (2.90) in fact is the linearized model of the single-machine infinite-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).
Solution of Eq. (2.90) is as follows:

DdðtÞ ¼ a0 e2Mt cos xNOF t þ b0


D
ð2:91Þ
r ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
where a0 and b0 are two constants and xNOF ¼ 1 D  4x0 K1 .
2 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
infinite-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 infinite-bus power system is determined by the coefficient of the
first-order derivative in the second-order differential equation in Eq. (2.91) M D.
At the angular oscillation frequency xs , the electric torque contributed from the
lower part of Heffron–Phillips model can be decomposed into two components

DTe ¼ Td Dx þ Ts Dd ð2:92Þ
44 2 A Single-Machine Infinite-Bus Power System …

Equation (2.89) becomes


   
D Td x0 K1 x0 Ts
s2 Dd þ þ sDd þ þ Dd ¼ 0 ð2:93Þ
M Mx0 M M

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 Heffron–Phillips model is as follows:

DTet ¼ Fdelta ðsÞDd þ Fpss ðsÞDupss ð2:94Þ

where Fdelta ðsÞ and Fpss ðsÞ 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 ¼ Fpss ðsÞDupss ð2:95Þ

Figure 2.15 shows that the PSS contributes the electric torque, DTpss , to the
electromechanical oscillation loop of generator. Obviously, Fpss ðsÞ is the transfer
function of forward path from the stabilizing signal of the PSS to the

ΔΤpss

K2

ΔE q’ 1 KA + Δupss
Td0's+K3 1+sTA

K6

Fig. 2.15 Forward path of stabilizing signal of the PSS


2.3 Damping Torque Analysis 45

electromechanical oscillation loop of generator in the Heffron–Phillips model.


From Fig. 2.15, it can be obtained that

1 KA
K3 þ sT0d0 1 þ sTA K
Fpss ðsÞ ¼ K2 ¼ K2  0
 A ð2:96Þ
1 þ K6 1
0
K A K3 þ sT d0 ð1 þ sTA Þ þ K6 KA
K3 þ sTd0 1 þ sTA

Hence, at the angular oscillation frequency, xs , the electric torque provided by


the PSS to the electromechanical oscillation loop is as follows:

pss ðjxs ÞDupss


DTpss ¼ 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 ¼ Kpss Dx, the electric torque
contributed from the PSS to the electromechanical oscillation loop of generator is as
follows:

pss ðjxs ÞDx ¼ Kpss Re½F


DTpss ¼ Kpss F pss ðjxs ÞDx þ jKpss Im½F
pss ðjxs ÞDx ð2:97Þ

where Re F pss ðjxs Þ and Im Fpss ðjxs Þ denote the real and imaginary part of
pss ðjxs Þ, respectively (this notation will be used throughout the book). From the
F
first equation in Eq. (2.41), it can have sDd ¼ x0 Dx, i.e.

jxs
Dx ¼ Dd ð2:98Þ
x0

By substituting Eqs. (2.98) into Eq. (2.97), it can be obtained that


pss ðjxs ÞDx  xs Kpss Im F


DTpss ¼ Kpss Re½F pss ðjxs Þ Dd ð2:99Þ
x0

pss ðjxs ÞDx, and the


Hence, the damping torque supplied by the PSS is Kpss Re½F
xs pss ðjxs ÞDd.
synchronizing torque is  x0 Kpss Im½F

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 ¼ aT238 Dx38 . From Eq. (2.78) or
Fig. 2.13, it can have
a22 a21
s2 Dd þ sDd þ x0 Dd þ DTe ¼ 0 ð2:100Þ
M M
46 2 A Single-Machine Infinite-Bus Power System …

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 ðsÞ ¼ aT238 ðsI  A33 Þ1 bpss3 ð2:101Þ

At the angular oscillation frequency, xs , the electric torque contribution from the
PSS is as follows:

DTpss ¼ aT238 ðjxs I  A33 Þ1 bpss3 Dupss ð2:102Þ

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 coefficient
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 3 2 32 3
sDd 0 x0 0 Dd
6 7 6 a21 a22 þ Dpss aT238 76 7
4 sDx 5 ¼ 4    M 54 Dx 5 ð2:103Þ
M M
Dx38 a138 a238 A33 Dx38

From Eqs. (2.70) and (2.103), it can have

i
@k @Ac
 Ti
¼w vi
@Dpss @Dpss
2 32 3
0 x0 0 vi1

@ 6 aT238 76 7
¼ w i1  i2
w  i3
w T
4  a21 
a22 þ Dpss
 M 54 vi2 5 ð2:104Þ
@Dpss M M
a138 a238 A33 vi3
 i2 
w vi2
¼
M

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
system. At the complex frequency k i , it is equal to the forward path of the PSS
multiplied by the sensitivity of the mode to the damping torque contribution.
2.3 Damping Torque Analysis 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 Phillips–Heffron model as follows.
Firstly, the effect of PSS is not considered, i.e. Dupss ¼ 0. From Fig. 2.10, it can
have
 2 
Ms þ Ds þ x0 K1 DdðsÞ ¼ x0 DTðsÞ
ð2:105Þ
DTðsÞ ¼ Fdelta ðsÞDdðsÞ

where Fdelta ðsÞ is the transfer function from DdðsÞ to DTðsÞ. Combining two
equations above gives
2

Ms þ Ds þ x0 K1 þ x0 Fdelta ðsÞ DdðsÞ ¼ 0 ð2:106Þ

Thus, characteristic equation of the system is as follows:

Ms2 þ Ds þ x0 K1 þ x0 Fdelta ðsÞ ¼ 0 ð2:107Þ

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 defines the damping of power
s ¼ n þ jxs . In the complex frequency domain, it
oscillation. Denote the mode as k s
should have

2 þ Dk
Mk  s þ x0 K1 þ x0 F s Þ ¼ 0
delta ðk ð2:108Þ
s

The second equation in Eq. (2.105) expressed in the complex frequency domain
is as follows:

s Þ ¼ F
DTðk s ÞDdðk
delta ðk s Þ ð2:109Þ

Also in the complex frequency domain, the first equation in Eq. (2.41) becomes

s Þ ¼ ns þ jxs s Þ ¼ ns Ddðk
s Þ þ j xs Ddðk
s Þ
Dxðk Ddðk ð2:110Þ
x0 x0 x0

Let the electric torque defined by Eq. (2.109) be decomposed as follows:

s Þ ¼ Ts1 Ddðk
DTðk s Þ þ Td1 Dxðk
s Þ ð2:111Þ
48 2 A Single-Machine Infinite-Bus Power System …

From Eqs. (2.109), (2.110), and (2.111), it can be obtained that

s ÞDdðk
delta ðk
F s Þ þ Td1 ns Ddðk
s Þ ¼ Ts1 Ddðk s Þ þ jT xs Ddðk
s Þ ð2:112Þ
x0 d1
x0

That is

F s Þ ¼ Ts1 þ Td1 ns þ jT xs
delta ðk ð2:113Þ
x0 d1
x0

From Eq. (2.113), it can have


(
delta ðk
Td1 ¼ xx0s Im½F s Þ
ð2:114Þ
Ts1 ¼ Re½F s Þ  Td1 ns
delta ðk
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 þ Dk
ðMk s þ x0 K1 ÞDdðk
s Þ ¼ x0 Ts1 Ddðk
s Þ  x0 Td1 Dxðks Þ
s
ð2:115Þ
¼ x0 Ts1 Ddðk s Ddðk
s Þ  Td1 k s Þ

Thus,

Mk s þ x0 K1 þ x0 Ts1 ¼ 0
2 þ ðD þ Td1 Þk ð2:116Þ
s

Solution of the above equation is as follows:

D þ Td1
ns ¼  ð2:117Þ
2M

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 ¼ x0 Dx
: 1 ð2:118Þ
D x ¼  ðDPt þ DDxÞ
M
2.3 Damping Torque Analysis 49

Without referring to any linearized model of single-machine infinite-bus power


system, assume DPt being comprised of contributions from Dd and stabilizing
signal of the PSS, Dupss . That is to let

DPt ¼ DPt ðDdÞ þ DPt ðDupss Þ


ð2:119Þ
DPt ðDdÞ ¼ Fdelta ðsÞDd; DPt ðDupss Þ ¼ Fpss ðsÞDupss

At a known angular oscillation frequency xs , if DPt is decomposed in the


complex frequency domain, then

DPt ðDdÞ ¼ F delta ðjxs ÞDd ¼ Cdelta Dd þ Ddelta Dx;


ð2:120Þ
pss ðjxs ÞDupss ¼ Cpss Dd þ Dpss Dx
DPt ðDupss Þ ¼ F

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 ¼ 0. When there is no PSS installed DPt ðDupss Þ ¼ 0 ,

DPt ¼ DPt ðDdÞ ¼ Cdelta Dd ð2:121Þ

The linearized DPt  Dd curve is line a–f 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 oper-
ating point moves down from the initial point a (Pt1 ; d1 ) along line a–f, it will stop at
point f (Pt20 ; d20 ) with area a–d–c being equal to area d–g–f. Obviously, in this case,
jPt1  Pt0 j ¼ jPt20  Pt0 j; jd1  d0 j ¼ 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 ðDupss Þ ¼ Dpss Dx ðassuming Dpss [ 0Þ ð2:122Þ

From Eqs. (2.120), (2.121), and (2.122), it should have

DPt ¼ DPt ðDdÞ þ DPt ðDupss Þ ¼ Cdelta Dd þ Dpss Dx; ð2:123Þ

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 a–f along curve
DPt ¼ Cdelta Dd þ Dpss Dx;. When the operating point stops moving, Dx ¼ 0. Thus,
it should stop on line a–f at point c (Pt2 ; d2 ) with area A1 being equal to area A2 .
50 2 A Single-Machine Infinite-Bus Power System …

Fig. 2.16 Graphical Pt ΔPt=CdeltaΔδ+Dpss Δω a


explanation of the DTA

ΔPt=CdeltaΔδ

A1
Pt0 g d e
A2
c
Direction of
f Δω < 0

δ
δ2' δ2 δ0 δ1
δ

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 ðsÞ, that is
2.3 Damping Torque Analysis 51

DuPSS ¼ Tpss ðsÞDx ð2:124Þ

From Eqs. (2.95) or (2.102) and above equation, it can have

DTpss ¼ Fpss ðsÞTpss ðsÞDx ð2:125Þ

At the angular oscillation frequency, xs , the decomposition of the electric torque


contributed by the PSS is as follows:

pss ðjxs ÞT
DTpss ¼ F  pss ðjxs ÞDx

¼ Re F pss ðjxs ÞT  pss ðjxs Þ Dx þ jIm½F


pss ðjxs ÞT
 pss ðjxs ÞDx

¼ Re½Fpss ðjxs ÞT  pss ðjxs ÞDx  xs Im½Fpss ðjxs ÞT pss ðjxs ÞDd ð2:126Þ
x0
¼ Tpssd Dx þ Tpsss Dd

The damping and synchronizing torque provided by the PSS is Tpssd Dx and
Tpsss Dd, respectively. In order to achieve the most efficient design, ideally the PSS
should provide only the damping torque, that is,

DTpss ¼ Dpss Dx; Dpss [ 0 ð2:127Þ

where Dpss is the coefficient 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 ðjxs ÞT
Dpss ¼ F  pss ðjxs Þ ð2:128Þ

According to Eq. (2.128), design of the PSS should set the phase of the PSS,
\T pss ðjxs Þ, to be equal to the minus phase of the forward path, \F
pss ðjxs Þ, that is to
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 ðjxs Þ ¼ Fpss \/; T
F  pss ðjxs Þ ¼ Tpss \c ð2:129Þ

The phase compensation method requires



Tpssd ¼ Fpss Tpss cosð/ þ cÞ ¼ Dpss
ð2:130Þ
Tpsss ¼ Fpss Tpss sinð/ þ cÞ ¼ 0
52 2 A Single-Machine Infinite-Bus Power System …

This can be achieved by setting


Dpss
c ¼ /; Tpss ¼ ð2:131Þ
Fpss

Often the PSS is constructed as a lead–lag block with its main part of transfer
function to be

ð1 þ sT2 Þ ð1 þ sT4 Þ ð1 þ sT2 Þ ð1 þ sT4 Þ


Tpss ðsÞ ¼ Kpss ¼ Kpss1 Kpss2 ð2:132Þ
ð1 þ sT1 Þ ð1 þ sT3 Þ ð1 þ sT1 Þ ð1 þ sT3 Þ

where Kpss ¼ Kpss1 Kpss2 . Parameters of the PSS then can be set to satisfy

ð1 þ jxs T2 Þ Dpss /
Kpss1 ¼ \
ð1 þ jxs T1 Þ Fpss 2
ð2:133Þ
ð1 þ jxs T4 Þ /
Kpss2 ¼ 1:0\
ð1 þ jxs T3 Þ 2

for the PSS to provide a positive damping torque Dpss Dx.


With the PSS installed, from Fig. 2.10, it can have

ðMs2 þ Ds þ x0 K1 ÞDdðsÞ ¼ x0 DTdelta ðsÞ  x0 DTpss ðsÞ


DTdelta ðsÞ ¼ Fdelta ðsÞDdðsÞ ð2:134Þ
DTpss ðsÞ ¼ Fpss ðsÞTpss ðs)DxðsÞ

c ¼ n þ 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 Þ ¼ F
DTdelta ðk c ðsÞDdðk
delta k c Þ ¼ C1 Ddðkc Þ þ D1 Dxðkc Þ
ð2:135Þ
DTpss ðkc Þ ¼ F c ÞT
pss ðk c ÞDxðk
 pss ðk c Þ ¼ Cpss Ddðk
c Þ þ Dpss Dxðk
c Þ

Taking the similar procedure of discussion in Sect. 2.3.2.1, it can easily prove
c Þ affects the damping of the electromechanical oscillation mode.
that Dpss Dxðk
Hence, the PSS can be designed to satisfy

c ÞTpss ðk
Fpss ðk c Þ ¼ Dpss ð2:136Þ

When the PSS is being designed, if k c ¼ n þ jxc is given, the PSS can be
c
designed by using the phase compensation method defined by Eq. (2.136) to move
c ¼ n þ 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 53

2.4 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 infinite-bus power system are as


follows:
1. Generator:

xd ¼ 1:18; xq ¼ 0:78; xad ¼ 1:0; xaq ¼ 0:6; xD ¼ 1:11; xQ ¼ 0:73; xf ¼ 1:13;


ra ¼ 0:005; rf ¼ 0:00075; rD ¼ 0:002; rQ ¼ 0:04; M ¼ 7 s; D ¼ 0; T0d0 ¼ 5 s

2. AVR: KA ¼ 100; TA ¼ 0:01 s


3. Transmission line: xt ¼ 0:15
4. Steady-state operating point: Pt0 ¼ 0:5; Vt0 ¼ 1:05; Vb0 ¼ 1:0
At the steady-state operating point, the complex power received at the infinite
busbar is as follows:
 
 t0  V
V  b0 

Vb0 ¼ Pt0 þ jQb0
jxt

where Qb0 is the reactive power received at the infinite busbar and * denotes the
conjugate of a complex phasor. By choosing V  b0 as the reference phasor, that is,
  ¼ Vb0 , from the above equation, it can be obtained that
 b0 ¼ V
V b0
2sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3
 2
Vb0 4 xt Pt0
Qb0 ¼ V2t0   Vb0 5 ¼ 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 ¼ Pt0  jQb0 ¼ 0:5  j0:3155


Vb0
54 2 A Single-Machine Infinite-Bus Power System …

Terminal voltage of generator is :

 t0 ¼ jxtIt0 þ Vb0 ¼ 1:0473 þ j0:075


V

At steady state, iD0 ¼ iQ0 ¼ 0; x0 ¼ 1, Eqs. (2.1) and (2.2) become

0 ¼ vtd0 þ ra id0 þ wq0


0 ¼ vtq0 þ ra iq0  wd0
0 ¼ vf0  rf if0
    
wd0 xd xad id0
¼
wf0 xad xf if0
wq0 ¼ xq iq0

Thus, according to Eq. (2.28)

vtd0 ¼ ra id0  wq0 ¼ ra id0 þ xq iq0


vtq0 ¼ ra iq0 þ wd0 ¼ ra iq0 þ xad if0  xd id0 ¼ ra iq0 þ Eq0  xd id0

Hence, in d–q coordinate of generator,

 t0 ¼ Vtd0 þ jVtq0 ¼ ra id0 þ xq iq0 þ jðra iq0 þ Eq0  xd id0 Þ


V
¼ ra ðid0 þ jiq0 Þ þ xq iq0  jxq id0 þ jðEq0  xd id0 þ xq id0 Þ

Let an imaginary electromotive force (EMF) be

EQ0 ¼ Eq0  xd id0 þ xq id0 ¼ Eq0  ðxd  xq Þid0

Thus,

 t0 ¼ ðra þ jxq Þðid0 þ jiq0 Þ þ E


V  Q0

Q
Thus, the q-axis of generator can be found by calculating the imaginary EMF E

 Q0 ¼ V
E  t0 þ ðra þ jxq ÞIt0 ¼ 1:2959 þ j0:4634 ¼ 1:38\19:68

Hence, d0 ¼ 19:68 . From the above equation about Eq0 and EQ0 , it can have

Eq0 ¼ EQ0  ðxq  xd Þid0 ¼ 1:1506


Eq0
if0 ¼ ¼ 1:5624
xad
2.4 Examples 55

Fig. 2.17 Phasor diagram for


computing initial values of q
state variables
EQ

[ra + j(x q − x d )]I t

Eq
jx d I t
v tq
jx t I t
Vt b
Vb

δ ϕ
It
vtd d

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 ¼ It0 sinðd0 þ u0 Þ ¼ 0:4654; iq0 ¼ It0 cosðd0 þ u0 Þ ¼ 0:3646


vd0 ¼ Vb0 sin d0 ¼ 0:3368; vq0 ¼ Vb0 cos d0 ¼ 0:9416

Because

 t0 ¼ vtd0 þ jvtq0 ¼ jxtIt0 þ Vb0 ¼ jxt ðid0 þ jiq0 Þ þ vd0 þ jvq0


V

thus

vtd0 ¼ vd0  xt iq0 ¼ 0:2820


vtq0 ¼ xt id0 þ vq0 ¼ 1:0114

From Eq. (2.23), it can have

Eq0
if0 ¼ ¼ 1:5624
xad
56 2 A Single-Machine Infinite-Bus Power System …

Because iD0 ¼ 0; iQ0 ¼ 0, thus according to Eq. (2.2), it can be obtained that
2 3 2 32 3 2 3
wd0 xd xad xad id0 1:0132
6 7 6 76 7 6 7
4 wf0 5 ¼ 4 xad xf xad 54 if0 5 ¼ 4 1:3001 5
wD0 xad xad xD iD0 1:0970
" #     
wq0 xq xaq iq0 0:2844
¼ ¼
wQ0 xaq xQ iQ0 0:2187

Denote
2 31 2 3 2 3
xd xad xad a11 a12 a13 4:2331 1:8311 2:1640
6 7 6 7 6 7
4 xad xf xad 5 ¼ 4 a21 a22 a23 5 ¼ 4 1:8311 5:1570 2:9963 5
xad xad xD a31 a32 a33 2:1640 2:9963 5:5498
 1    
xq xaq b11 b12 3:6842 2:8653
¼ ¼
xaq xQ b13 b14 2:8653 3:7249

From Eqs. (2.12) and (2.13), it can have


2 3
0 x0 0 0 0 0 0 0
6 b V b V 7
60 MD 0  a11M
Vtd0  11M tq0  a12M
Vtd0  a13M
Vtd0  12M tq0 7
6 7
60  T1 7
6 0 0 0 0 0 0 7
6 A 7
60 x w a11 x0 ra x0 x a12 x0 ra a13 x0 ra 7
Agdq ¼6
6
0 q0 0 0 7
7
6 0 x0 wd0 0 x0 x b11 x0 ra 0 0 b12 x0 ra 7
6 7
60 x0 a21 x0 rf a22 x0 rf a23 x0 rf 7
6 0 0 0 7
6 7
40 0 0 a31 x0 rD 0 a32 x0 rD a33 x0 rD 0 5
0 0 0 0 b21 x0 rQ 0 0 b22 x0 rQ
2 3
0 314:16 0 0 0 0 0 0
60 0:0738 0:0872 0:4140 7
6 0 0 0:1706 0:5037 7
6 7
60 0 100 0 0 0 0 0 7
6 7
6 0 89:3377 6:6494 7
6 0 314:16 2:8762 3:3992 0 7
¼6 7
6 0 318:3167 0 314:16 5:4760 0 0 4:5008 7
6 7
60 1:2151 7
6 0 314:16 0:4314 0 0:7060 0 7
6 7
40 0 0 1:35797 0 1:8826 3:4871 0 5
0 0 0 0 36:0068 0 0 46:8088
2.4 Examples 57

2 0 0 3 2 3
0 0
6  id0 iq0 7 6
6 M 7 0:0665 0:0521 7
6 M 7 6 7
6 KA Vtd0 KA Vtq0 7 6 2686:1 9632:5 7
6
7
6 T V T V 7 6 7
6 t0 7
6 7
6 x 7 6 314:1593
A t0 A

6 0 7
0 7
Bgdq ¼6 0 ¼
7 6 6 7
6 x 7 0 314:1593 7
6 0 0 7 6 6 7
6 7 6 7
6 0 0 7 6 0 0 7
6 7 4 7
4 5 0 0 5
0 0
0 0 0 0
 
0 0 0 a11 0 a12 a13 0
Cgdq ¼
0 0 0 0 b11 0 0 b12
 
0 0 0 4:2331 0 1:8331 2:1640 0
¼
0 0 0 0 3:4862 0 0 2:8653
   
0 xt 0 0:15
Fdq1 ¼ ¼
xt 0 0:15 0
   
Vb cos d0 0 0:9416 0
Fdq2 ¼ ¼
Vb sin d0 0 0:3367 0

Thus, state matrix and control vector of linearized state-space model are obtained
to be

Agcdq ¼ Agdq þ Bgdq Fdq1 Cgdq þ Bgdq Fdq2


2 3
0 314:16 0 0 0 0 0 0
6 0:045 0:09 0:104 0:38 7
6 0 0 0:20 0:47 7
6 7
6 714:3 0 100 6116:3 1404:6 2646 3127 1154:5 7
6 7
6 295:8 89:34 6:64 135:03 7
6 0 478:4 2:88 3:39 7
¼6 7
6 105:78 318:3 0 513:64 5:48 86:29 101:9 4:50 7
6 7
6 1:22 7
6 0 0 314:2 0:43 0 0:70 0 7
6 7
4 0 0 0 1:36 0 1:88 3:48 0 5
00 0 0 36:0 0 0 46:8
2 2 3 3
0 0
6 0 7 6 0 7
6 7 6 7
6 KA 7 6 7
6 T 7 6 10000 7
6 A7 6 7
6 0 7 6 0 7
6 7 6 7
bpss ¼6 7¼6 7
6 0 7 6 0 7
6 7 6 7
6 0 7 6 0 7
6 7 6 7
6 7 6 7
4 0 5 4 0 5
0 0
58 2 A Single-Machine Infinite-Bus Power System …

Eigenvalues of above state matrix are computed to be


 1 ;k
k 2 ¼ 24:14  j971:56
 3 ;k
k 4 ¼ 36:33  j359:63
k5 ¼ 35:57
k6 ; k
7 ¼ 0:71  j8:44
k8 ¼ 5:71

According to Eq. (2.63), the oscillation frequency f i (Hz) and damping fi for the
electromechanical oscillation mode, k 6 and k7 , are as follows:
xi
fi ¼ ¼ 1:34 Hz
2p
ni
fi ¼  qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ 0:0838; i ¼ 6; 7
ni þ x2i
2

2.4.1.2 Heffron–Phillips 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
x2
x0d ¼ xd  ad ¼ 0:2951
x
  f
E0q0 0
¼ EQ0  xq  xd itd0 ¼ 1:1506

From Eq. (2.40), it can be obtained that

K1 ¼ 1:5248; K2 ¼ 0:7602; K3 ¼ 2:9885;


K4 ¼ 0:6727; K5 ¼ 0:0027; K6 ¼ 0:3245

State equation of (2.41) can be obtained to be


2 : 3 2 32 3 2 3
Dd 0 314:16 0 0 Dd 0
6 : 7
6 D x0 7 6 0:2178 0 0:1086 0 7 6 7 6 7
76 Dx0 7 þ 6 0 7Dupss
6 _ 7¼6
4 DEq 5 4 0:1346 0 0:5978 0:2 54 DEq 5 4 0 5
0 0
DE_ 26:4729 0 3245 100 DEfd 10000
fd

Eigenvalues of state matrix are as follows:

2;3 ¼ 0:0114  8:2610; k4 ¼ 7:6008


k1 ¼ 92:9741; k
2.4 Examples 59

According to Eq. (2.63), oscillation frequency f i (Hz) and damping fi for the
electromechanical oscillation mode, k2;3 , are as follows:

xi
fi ¼ ¼ 1:315 Hz
2p
ni
fi ¼  qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ 0:0014; i ¼ 2; 3
ni þ x2i
2

2.4.2 Modal Analysis of Example Power System

Heffron–Phillips 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 3 2 3
8:4781  106 0:9014
6 2:5091  106 7 6 7
6 7 6 j0:0237 7
v1 ¼ 6 7; v2 ¼ 6 7
4 0:0022 5 4 0:0050 þ j0:0050 5
1 0:3868  j0:1930
2 3 2 3
0:9014 0:0077
6 j0:0237 7 6 0:0002 7
6 7 6 7

v3 ¼ 6 7; v4 ¼ 6 7
4 0:0050  j0:0050 5 4 0:0284 5
0:3868 þ j0:1930 0:9996

Hence,

 ¼ ½ v1 
V v2  v3 v4 
2 3
8:4781  106 0:9014 0:9014 0:0077
6 2:5091  106 j0:0237 7
6 j0:0237 0:0002 7
¼6 7
4 0:0022 0:0050 þ j0:0050 0:0050  j0:0050 0:0284 5
1 0:3868  j0:1930 0:3868 þ j0:1930 0:9996
60 2 A Single-Machine Infinite-Bus Power System …

From Eq. (2.53),

 T ¼ ½ w1  1
W 2
w 3
w w4 T ¼ V
2 3
0:2524 0:5554  j0:0001 0:5554 þ j0:0001 0:1773
6 0:8529 0:0312  j21:1226 0:0312 þ j21:1226 7:3269 7
6 7
¼6 7
4 38:0214 0:1626 þ j0:1475 0:1626  j0:1475 38:2206 5
1:0823 0:0003 þ j0:0003 0:0003  j0:0003 0:0827

Modal decomposition thus is obtained to be

szi ¼ ki zi ; þ wTi bo u, i ¼ 1; 2; 3; 4
 
Without considering the PSS Dupss ¼ 0 , solution of modal decomposition is
obtained to be

zi ðt) ¼ zi ð0Þeki t ; i ¼ 1; 2; 3; 4

From Eq. (2.55), it can have

Zð0Þ ¼ V1 Xð0Þ ¼ WT Xð0Þ

Time response of state variables can be written as follows:

X
4 X
4
xk ðt) ¼ vki zi ð0Þeki t ¼ vki zi ð0Þeni t ½cos xi t þ j sin xi t
i¼1 i¼1

It can be seen that when time approaches infinity ðt ! 1Þ,

X
4 X
4
lim xk ðt) ¼ lim vki zi ð0Þeki t ¼ lim vki zi ð0Þeni t ½cos xi t þ j sin xi t
t!1 t!1 t!1
i¼1 i¼1
92:9741t
¼ lim fe vk1 z1 ð0Þ þ e0:0114t vk2 z2 ð0Þ½cosð8:2610tÞ þ j sinð8:2610tÞ
t!1
þ e0:0114t vk3 z3 ð0Þ½cosð8:2610tÞ  j sinð8:2610tÞ þ e7:6008t vk4 z4 ð0Þg ¼ 0


T
Since XðtÞ ¼ DdðtÞ DxðtÞ DE0q ðtÞ DE0fd ðtÞ and

DdðtÞ ¼ dðtÞ  d0 ; DxðtÞ ¼ xðtÞ  x0 ;


DE0q ðtÞ ¼ E0q ðtÞ  E0q0 ðtÞ; DE0fd ðtÞ ¼ E0fd ðtÞ  E0fd0 ðtÞ
2.4 Examples 61

lim xk ðt) ¼ 0; k ¼ 1; 2; 3; 4 should give


t!1

lim dðtÞ ¼ d0 ; lim xðtÞ ¼ x0 ; lim E0q ðtÞ ¼ E0q0 ðtÞ; lim E0fd ðtÞ ¼ E0fd0 ðtÞ
t!1 t!1 t!1 t!1

All the state variables return to their initial points Xð0Þ, 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 ¼ D ¼ 0:2178
KM M
1
0
" #  
0 0
b0 ¼ K ¼ cT0 ¼ ½ 1 0 
M 2
0:1086

1;2 ¼ j8:2725 with the corresponding


Its eigenvalues are calculated to be k
eigenvectors to be
   
0:9997 0:9997
v1 ¼ ; v2 ¼
j0:0263 j0:0263

Thus

 T ¼ ½w
W 1 w  1 ¼ ½ v1 v2 1
 2 T ¼ V
   T
0:9997 0:9997 1 0:5002 0:5002
¼¼ ¼
j0:0263 j0:0263 j18:9947 j18:9947

Hence, the controllability and observability index, respectively, can be obtained


to be
 
0
 T1 b0
w ¼ ½ 0:5002 j18:9914  ¼ j2:0628
0:1086
 
0
 T2 b0
w ¼ ½ 0:5002 j18:9914  ¼ j2:0628
0:1086
 
0:9997
cT0 
v1 ¼ ½1 0 ¼ 0:9997
j0:0263
 
0:9997
cT0 
v2 ¼ ½1 0 ¼ 0:9997
j0:0263
62 2 A Single-Machine Infinite-Bus Power System …

The residue is as follows:

1 ¼ w
R  T1 b0  cT0 v1 ¼ j2:0622
2 ¼ w
R  T2 b0  cT0 v2 ¼ j2:0622

From Eq. (2.77), when a ¼ KA , it can have

ðTA ki þ 1ÞK4 þ KA K5


i
@k @  0 i þ 1Þ þ KA K6
ðki Tdo þ K3 ÞðTA k
¼ Ri
@KA @KA
i T0 þ K3 ÞðTA k
K5 ½ðk i þ 1Þ þ KA K6   ½ðTA k
i þ 1ÞK4 þ KA K5 K6
¼ Ri do
½ðki T0 þ K3 ÞðTA ki þ 1Þ þ KA K6 2
do
¼ 0:00018158  j0:00005839 i ¼ 1; 2

When a ¼ TA it can have



i
@k @  0ðTA ki þ 1ÞK4 þ KA K5
ðki Tdo þ K3 ÞðTA ki þ 1Þ þ KA K6
¼ Ri
@TA @TA
 i K4 ½ðk
k  i T0 þ K3 ÞðTA k i þ 1Þ þ KA K6   ½ðTA k i þ 1ÞK4 þ KA K5 ðk
 i T0 þ K3 Þk
i
¼ Ri do do
½ðki T þ K3 ÞðTA k
0  i þ 1Þ þ KA K6  2
do
¼ 0:0357  j0:1533 i ¼ 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 detri-
mental 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 ¼ 50 to KA ¼ 100 with TA ¼ 0:01 and TA increases from TA ¼ 0:01 to TA ¼
0:1 with KA ¼ 100. From Fig. 2.18, it can be seen that although at the point
KA ¼ 100; TA ¼ 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 ¼ 100; TA ¼ 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 ¼ 0:0114  j8:2611, to a
target position on the complex plane, kc ¼ 0:9  j8:2611, which is damping over
10 %.
Let the transfer function of the PSS be (see Eq. (2.88)) with T ¼ 0:1 s

ð1 þ saTÞ2
Tpss ðsÞ ¼ Kpss
ð1 þ sTÞ2
64 2 A Single-Machine Infinite-Bus Power System …

State-space model of example power system with the PSS to be installed is


2 3 2 32 3 2 3
Dd_ 0 314:16 0 0 Dd 0
6 7 6 76 Dx 7 6 7
6 Dx_ 7 6 0:2178 0 0:1086 76 0 7 6 7 0
6 7¼6 76 7 þ 6 7Dupss
6 DE_ 0 7 4 0:1346 0:5978 0:2 54 DE0q 5 4 0 5
4 q 5 0
DE_
0
26:4729 0 3245 100 DE0fd 10000
fd
2 3
Dd
6 Dx 7
6 7
Dx ¼ ½ 0 1 0 0 6 7
4 DE0q 5
DE0fd
Dupss ¼ Tpss ðsÞDx

Thus, the transfer function of open-loop system can be obtained to be

238:8s
GðsÞ ¼ cT ðsI  AÞ1 b ¼
1:1s4 þ 110:6s3 þ 854:5s2 þ 7563s þ 53020

From the characteristic equation of closed-loop control system of Eq. (2.85), for
the electromechanical oscillation mode k c , it should have

c Þ ¼ 1 1
Tpss ðk c Þ ¼ 0:0753  j0:0996
Gðk

That is
c aTÞ2
ð1 þ k 1
Kpss ¼

ð1 þ kc TÞ 2 0:0753  j0:0996

By solving the above equations, parameters of the PSS can be obtained to be



Kpss ¼ 2:6151
a ¼ 2:4256

In order to establish the state-space model of closed-loop system with the PSS
installed, let

ð1 þ saTÞ
Dx1 ¼ Dx
ð1 þ sTÞ
ð1 þ saTÞ
Dupss ¼ Kpss Dx1
ð1 þ sTÞ
2.4 Examples 65

Thus, state equation of the PSS is as follows:

1 1
sDx1 ¼ Dx þ a  sDx  Dx1
T  T 
K1 1 D K2 1
¼  aDd þ  a Dx  aDE0q  Dx1
M T M M T
1 1
sDupss ¼ Kpss Dx1  Dupss þ Kpss a  sDx1
T T
K1 2 1 D K2 0
¼ Kpss a Dd þ Kpss að  aÞDx  Kpss a2 DEq
M T M M
1 1
þ ð1  aÞ Kpss Dx1  Dupss
T T

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 : 32 32 3
Dd 0 314:159 0 0 0 0 Dd
6 : 7 6
6 Dx 7 6 0:218 0 0:109 0 0 0 7 6
76 Dx0 7
7
6 0 7 6
6 DE_ q 7 6 0:134 0 0:597 0:200 0 0 7 76
6 DEq 7
7
6 0 7 ¼ 6 26:582 10000 7 6 0 7
6 DE_ 7 6
6 fd 7
0 3245:044 100 0 76 DEfd 7
4 Dx_ 1 5 4 0:528 10 0:263 0 10 0 54 Dx1 5
Du_ pss 3:352 63:432 1:671 0 37:281 10 Dupss

Eigenvalues of state matrix are calculated to be

1 ¼ 93:4535
k
2;3 ¼ 0:8995  j5:2621
k
4;5 ¼ 9:4136  j5:0918
k
k6 ¼ 6:5176

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 2 A Single-Machine Infinite-Bus Power System …

rotor angle δ (degree)


50
without PSS with PSS installed

40

30

20

10

-10
0 1 2 3 4 5 6 7 8 9 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

Heffron–Phillips 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:

K5 1 KA
sT0do þ K3 sTA þ 1
DTavr ¼ K2 Dd
1 þ 0 K6 sTKAþ 1
sTdo þ K3 A
K2 K5 KA
¼ Dd ¼ Favr ðsÞDd
K6 KA þ ðsT0do þ K3 ÞðsTA þ 1Þ

At the complex frequency of the electromechanical oscillation with


s ¼ n þ jxs ¼ k1 ¼ 0:0114 þ j8:2610, let the decomposition of the electric
k s
torque provided by the AVR be
2.4 Examples 67

     
s ¼ Tsavr Dd k
DTavr k s þ Tdavr Dx k
s

Because

s Þ ¼ ns þ jxs s Þ ¼ ns Ddðk
s Þ þ j xs Ddðk
s Þ
Dxðk Ddðk
x0 x0 x0

thus

DTavr ðk s Þ þ Tdavr ns Ddðk


s Þ ¼ Tsavr Ddðk s Þ þ jT xs s Þ
Ddðk
x0 davr
x0

That is
xs s Þ
Tdavr ¼ ImFavr ðk
x0

Since

s Þ ¼  K2 K5 KA
Favr ðk  s TA þ 1Þ
K6 KA þ ðks T0do þ K3 Þðk
K2 K5 KA
¼
K6 KA þ K3 þ ðK3 TA þ T0d0 Þk 2
s þ T0 TA k
d0 s
K2 K5 KA
¼
K6 KA þ K3 þ ðK3 TA þ T0d0 Þðns þ jxs Þ þ T0d0 TA ðns þ jxs Þ2
K2 K5 KA K2 K5 KA
¼ ¼ 2 ða  jbÞ ¼ 0:0024 þ j0:0031
a þ jb a þ b2

where
   
a ¼ K6 KA þ K3 þ K3 TA þ T0d0 ns þ T0d0 TA n2s  x2s ¼ 31:9695
 
b ¼ K3 TA þ T0d0 xs þ 2T0d0 TA ns xs ¼ 41:5426

thus it can be obtained

x0 s Þ ¼ x0 K2 K5 KA b
Tdavr ¼ ImFavr ðk
xs xs a2 þ b2
¼ 0:0988
68 2 A Single-Machine Infinite-Bus Power System …

Since

@a @a
 
@ KA
¼ K6 ¼ 0:3245; @ TA
¼ K3 ns þ T0d0 n2s  x2s ¼ 341:2561
@b
@ KA
¼ 0; @b
@ TA
¼ K3 xs þ 2T0d0 ns xs ¼ 23:7488

sensitivity of the damping torque provided by the AVR to its parameters can be
obtained to be
h   i
@Tdavr x0 K2 K5 b þ KA @@Kb a2 þ b2  KA b 2a @@Ka þ 2b @@Kb
¼ A A A
¼ 0:0003
@KA xs ða2 þ b2 Þ2
h i
@b @a @b
x0 K2 K5 KA @ TA ða þ b Þ  b 2a @ TA þ 2b @ TA
2 2
@Tdavr
¼ ¼ 0:9055
@TA xs ða2 þ b2 Þ2

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 ¼ Dpss Dx; Dpss ¼ 15.
From Eq. (2.96), the forward path of stabilizing signal of the PSS can be obtained to
be ðjxs ¼ j8:44Þ
KA
Fpss ðjxs Þ ¼ K2 0
ðK3 þ jxs Td0 Þð1 þ jxs TA Þ þ K6 KA
¼ 0:8598  j1:1451 ¼ 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:

ð1 þ sT2 Þ ð1 þ sT4 Þ
Tpss ðsÞ ¼ K1 K2 with T1 ¼ 0:09 s; T3 ¼ 0:09 s
ð1 þ sT1 Þ ð1 þ sT3 Þ
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 ¼ Kpss1 Kpss2 ¼ 3:2271; T2 ¼ 0:2405 s; T4 ¼ 0:2405 s

Let
ð1 þ sT4 Þ
Dx1 ¼ Kpss2 Dx ¼ ð9:3706 þ j4:6819ÞDx
ð1 þ sT3 Þ
ð1 þ sT2 Þ
Dupss ¼ Kpss1 Dx1 ¼ ð0:8946 þ j0:4470ÞDx1
ð1 þ sT1 Þ

Thus, state-space realization of the PSS is as follows:

1 Kpss2
sDx1 ¼  Dx1 þ ðT4 sDx þ DxÞ
T3 T3
 
1 Kpss2 T4
¼  Dx1 þ ðK1 Dd  K2 DE0q  DDxÞ þ DxÞ
T3 T3 M
 
Kpss2 T4 K1 Kpss2 T4 Kpss2 T4 K2 0 1
¼ Dd þ 1  D Dx  DEq  Dx1
T3 M T3 M T3 M T3
¼ 0:3230Dd þ 6:1672Dx  0:1224DE0q  11:1111Dx1
  
1 Kpss1 T2 Kpss1 Kpss1 T2 Kpss2 T4 K1
sDupss ¼  Dupss þ  Dx1 þ  Dd
T1 T1 T3 T1 T1 T3 M
  
Kpss2 T4 K2 0 Kpss2 T4
 DEq þ 1  D Dx
T3 M T3 M
 
Kpss T2 T4 K1 Kpss T2 T4 Kpss T2 T4 K2 0
¼ Dd þ 1  D Dx  DEq
T1 T3 M T1 T3 M T1 T3 M
 
Kpss1 T2 Kpss1 1
+  Dx1  Dupss
T1 T3 T1 T1
¼ 5:0182Dd þ 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 2 A Single-Machine Infinite-Bus Power System …

2 0 x0 0 0 0 0 3
6  K1  D K 0 7
6 0 0 7
2

6 M M M 7
6 K 7
6  K0 4 0  03  1
0 0 7
6 Td0 Td0 T0d0 7
6 7
A¼6 KA 7
6  KTA K5 0  KTA K6 K A
0 TA 7
6 A
A TA 7
6 7
6  Kpss2 T4 K1 Kpss2
1  TTM  KT 20 T4 K2
K2 0  T1 0 7
6 7
4

4 T 3M T 3M
5

3 3 3

Kpss1
T 2 T 4 K1
T1 T3 M K PSS
T2 1  T4 D K
T1 T3 M PSS T 2 T 4 K2
T1 T3 M KPSS 0 T1 1  T2
T3 T1
1
2 3
0 314:16 0 0 0 0
6 0:2178 0:1086 7
6 0 0 0 0 7
6 7
6 0:1345 0 0:5977 0:2 0 0 7
¼6
6 26:5821
7
6 0 3245 100 0 10000 7
7
6 7
4 0:3230 6:1672 0:1224 0 11:1111 0 5
5:0182 95:8026 2:5018 0 108:0008 11:1111

Eigenvalues of state matrix are calculated to be

k1 ¼ 93:6838
2;3 ¼ 1:2125  j8:0051
k
k4;5 ¼ 8:7171  j6:2080
k6 ¼ 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 1
K4 1
K5 KA
1
0 0
B sT þ K3 sT sTA þ 1C
þ K3
Fdelta ðsÞ ¼ K2 @ þ
KA A
do do
K
1 þ 0 6 sTKAþ 1 1þ 0 K 6
sTdo þ K3 A sTdo þ K3 sTA þ 1

K2 ½K4 ðsTA þ 1Þ þ K5 KA 
¼  
K6 KA þ sT0do þ K3 ðsTA þ 1Þ

At the complex frequency of electromechanical oscillation with


s ¼ n þ jxs ¼ k1 ¼ 0:0114 þ j8:2610,
k s
 
delta ks ¼ 0:0042 þ j0:0042
F

According to Eq. (2.114)


(

Td1 ¼ xx0s Im F s Þ ¼ 0:1591


delta ðk

Ts1 ¼ Re F s Þ  Td1 ns ¼ 0:0042


delta ðk
x0

Substituting the above result into Eq. (2.116), it can have

2 þ 0:1553k
7k s þ 477:7128 ¼ 0
s

Solution of above equation is ks ¼ 0:0114 þ j8:2610. It is the electrome-


chanical oscillation mode of example power system without the PSS being
installed.
From Fig. 2.10, it can have

s ÞDd ¼ K1 Dd þ Ts1 Dd þ Td1 Dx ¼ 1:5207Dd þ 0:1553Dx


DPt ¼ K1 Dd þ Fdelta ðk

This is the first 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 2 A Single-Machine Infinite-Bus Power System …

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 first cycle of power oscillation starting from its first
peak. Dashed curve is the case without the PSS installed.
With the PSS designed in Sect. 2.4.3.2 being installed,

c ÞDd þ F
DPt ¼ K1 Dd þ Fdelta ðk c ÞT
pss ðk c ÞDxðk
 pss ðk c Þ
¼ K1 Dd þ Ts1 Dd þ Td1 Dx þ Dpss Dxðk c Þ
¼ 1:5207Dd þ 15:1553Dx

where kc ¼ 1:2279 þ 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
confirms the graphical explanation of the DTA illustrated in Fig. 2.16.
2.4 Examples 73

2.4.4 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 Heffron–Phillips 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 3 32 3 2 3
Dd 0 314:16 0 0 Dd 0
6 : 7 6
6 D x0 7 6 0:2178 0 0:1086 0 7 6 7 6 7
76 Dx0 7 þ 6 0 7Dupss
6 _ 7¼4 5 4 5 4
4 DEq 5 0:1345 0 0:5977 0:2 DEq 0 5
0 0
DE_ 26:5821 0 3245 100 DEfd 10000
fd 2 3
Dd
6 Dx 7
Dx ¼ ½ 0 1 0 0 6 4 DE0 5
7
q
DE0fd

The electromechanical oscillation modes are as follows:

k2;3 ¼ 0:0114  j8:2610

In Sect. 2.4.2.1, matrix formed by right eigenvectors is as follows:

 ¼ ½ v1
V v2 v3 v4 

Thus, the right eigenvectors related to the electromechanical oscillation modes


are as follows:
2 3 2 3
0:9014 0:9014
6 j0:0237 7 6 j0:0237 7
2 ¼ 6
v 7 6 7
4 0:0050 + j0:0050 5 v3 ¼ 4 0:0050  j0:0050 5
0:3868  j0:1930 0:3868 + j0:1930
74 2 A Single-Machine Infinite-Bus Power System …

Matrix formed by left eigenvectors is as follows:

 T ¼ ½ w1 1
W 2
w 3
w w 4 T ¼ V

The left eigenvectors related to the electromechanical oscillation modes are as


follows:

 T2 ¼ ½ 0:5554  j0:0001 0:0312  j21:1226 0:1626 þ j0:1475 0:0003 þ j0:0003 


w
 T3 ¼ ½ 0:5554 þ j0:0001 0:0312 þ 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 ¼ w
R  T2 bo cTo 
v2
¼ ½ 0:5554  j0:0001 0:0312  j21:1226 0:1626 þ j0:1476 0:0003 þ j0:0003 
2 3 2 3
0 0:9014
6 0 7 6 7
6 7 6 j0:0237 7
6 7½ 0 1 0 0 6 7
4 0 5 4 0:0050 + j0:0050 5
10000 0:3868  j0:1930
¼ 0:0633 þ j0:0822

3 ¼ w
R  T3 bo cTo 
v3
¼ ½ 0:5554 þ j0:0001 0:0312 þ j21:1226 0:1626  j0:1475 0:0003  j0:0003 
2 3 2 3
0 0:9014
6 0 7 6 j0:0237 7
6 7 6 7
6 7½ 0 1 0 0 6 7
4 0 5 4 0:0050  j0:0050 5
10000 0:3868 + j0:1930
¼ 0:0633  j0:0822

From Eq. (2.104), sensitivity of the electromechanical oscillation modes to the


coefficient of damping torque provided by the PSS can be computed to be

2
@k  22 v22
w ð0:0312  j21:1226Þðj0:0237Þ
¼ ¼ ¼ 0:0715
@Dpss M 7
3
@k  32 v32
w ð0:0312 þ j21:1226Þðj0:0237Þ
¼ ¼ ¼ 0:0715
@Dpss M 7
2.4 Examples 75

At the complex frequency k 2;3 ¼ 0:0114  j8:2610, the forward path can be
calculated from Eq. (2.96) as

2;3 Þ ¼ K2 KA
Fpss ðk
2;3 T0 Þð1 þ k2;3 TA Þ þ K6 KA
ðK3 þ k d0
0:7602  100
¼
½2:9885 þ ð0:0114  j8:2611Þ  5½1 þ ð0:0114  j8:2611Þ  0:01 þ 0:3245  100
¼  0:8845  j1:1493

Hence,

@k2;3
R2;3 ¼  Fpss ðk2;3 Þ ¼ ð0:0715Þ  ð0:8845  j1:1493Þ
@Dpss
¼ 0:0633  j0:0822

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 coeffi-
cient 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 3
0 314:16 0 0 0 0 0 0
6 0:045 0:09 0:104 0:39 7
6 0 0 0:20 0:47 7
6 7
6 714:29 0 100 6116:3 1404:6 2646 3127 1154:5 7
6 7
6 295:81 89:35 6:65 135:03 7
6 0 478:4 2:88 3:40 7
Agcdq ¼6 7
6 105:79 318:31 0 513:64 5:48 86:29 101:98 4:50 7
6 7
6 1:22 7
6 0 0 314:16 0:43 0 0:71 0 7
6 7
4 0 0 0 1:36 0 1:88 3:49 0 5
0 0 0 0 36:0 0 0 46:8
2 3
0
6 0 7
6 7
6 7
6 10000 7
6 7
6 0 7
6 7
bpss ¼6 7
6 0 7
6 7
6 0 7
6 7
6 7
4 0 5
0
76 2 A Single-Machine Infinite-Bus Power System …

Eigenvalues of state matrix are as follows:

1;2 ¼ 24:12  j971:56


k
3;4 ¼ 36:32  j359:63
k
k5 ¼ 35:57
6;7 ¼ 0:71  j8:44
k
k8 ¼ 5:71

The pair of electromechanical oscillation modes are as follows:

k6;7 ¼ 0:71  j8:44

For each of eigenvalues, right eigenvector is calculated to be


2 3 2 3
0:000005  j0:000011 0:000005 þ j0:000011
6 0:000032 þ j0:000016 7 6 0:000032  j0:000016 7
6 7 6 7
6 7 6 7
6 0:9507 7 6 0:9507 7
6 7 6 7
6 0:0001 þ j0:0182 7 6 0:0001  j0:0182 7
6 7 6 7
v1 ¼ 6 7v2 ¼ 6 7
6 0:0369 þ j0:0014 7 6 0:0369  j0:0014 7
6 7 6 7
6 0:0072  j0:3072 7 6 0:0072 þ j0:3072 7
6 7 6 7
6 7 6 7
4 0:000569 þ j0:000026 5 4 0:000569  j0:000026 5
j0:0014 j0:0014
2 3 2 3 2 3
0:000215  j0:000026 0:000215 þ j0:000026 0:0527
6 0:000055  j0:000244 7 6 0:000055 þ j0:000244 7 6 0:006 7
6 7 6 7 6 7
6 7 6 7 6 7
6 0:7251 7 6 0:7251 7 6 0:0539 7
6 7 6 7 6 7
6 0:0163  j0:2196 7 6 0:0163 þ j0:2196 7 6 0:1055 7
6 7 6 7 6 7
v3 ¼ 6 7v4 ¼ 6 7 v5 ¼ 6 7
6 0:1654 þ j0:025 7 6 0:1654  j0:025 7 6 0:2563 7
6 7 6 7 6 7
6 0:0615  j0:6274 7 6 0:0615 þ j0:6274 7 6 0:4925 7
6 7 6 7 6 7
6 7 6 7 6 7
4 0:0041 þ j0:0006 5 4 0:0041  j0:0006 5 4 0:0334 5
0:003  j0:0165 0:003 þ j0:0165 0:8209
2 3 2 3 2 3
0:7004 0:7004 0:000104
6 0:0016 þ j0:0188 7 6 0:0016  j0:0188 7 6 0:000002 7
6 7 6 7 6 7
6 7 6 7 6 7
6 0:000061  j0:000914 7 6 0:000061 þ j0:000914 7 6 0:0095 7
6 7 6 7 6 7
6 0:1482 þ j0:0043 7 6 0:1482  j0:0043 7 6 0:000116 7
6 7 6 7 6 7
v6 ¼ 6 7v7 ¼ 6 7 v8 ¼ 6 7
6 0:5516 þ j0:0283 7 6 0:5516  j0:0283 7 6 0:000082 7
6 7 6 7 6 7
6 0:0317 þ j0:0041 7 6 0:0317  j0:0041 7 6 0:7636 7
6 7 6 7 6 7
6 7 6 7 6 7
4 0:0077 þ j0:0284 5 4 0:0077  j0:0284 5 4 0:6456 5
0:413 þ j0:0977 0:413  j0:0977 0:000072
2.4 Examples 77

They form the following matrix

 ¼ ½ v1
V v2 ... v8 

 ¼ ½w
T 1
 ; left eigenvectors corresponding to
As W 1 w  2 . . . w8 T ¼ V
eigenvalues are calculated to be
2 3T 2 3T
1:2419  j0:4483 1:2419 þ j0:4483
6 0:0486  j0:262 7 6 0:0486 þ j0:262 7
6 7 6 7
6 0:4494 þ j0:0412 7 6 0:4494  j0:0412 7
6 7 6 7
6 0:2153  j3:8659 7 T 6 0:2153 þ j3:8659 7
 T1
w ¼6 7 w ¼ 6
6 1:9686 þ j0:7712 7 2 6 1:9686  j0:7712 7
7
6 7 6 7
6 0:0188 þ j1:3997 7 6 0:0188  j1:3997 7
6 7 6 7
4 0:0288 þ j1:6542 5 4 0:0288  j1:6542 5
0:6025  j0:5408 0:6025 þ j0:5408
2 3T 2 3T 2 3T
1:6552 þ j0:1633 1:6552  j0:1633 0:7411
6 0:1388 þ j0:9214 7 6 0:1388  j0:9214 7 6 6:8756 7
6 7 6 7 6 7
6 0:1003  j0:0038 7 6 0:1003 þ j0:0038 7 6 0:000219 7
6 7 6 7 6 7
6 0:2788 þ j1:9563 7 T 6 0:2788  j1:9563 7 T 6 0:0934 7
 T3 6
¼6 7 6
4 ¼ 6 7 6 7
w 7 w 7 w5 ¼ 6 0:0106 7
6 2:5805  j0:126 7 6 2:5805 þ j0:126 7 6 7
6 0:0246 þ j0:1141 7 6 0:0246  j0:1141 7 6 0:000045 7
6 7 6 7 6 7
4 0:0277 þ j0:1353 5 4 0:0277  j0:1353 5 4 0:00006 5
0:7412  j0:2714 0:7412 þ j0:2714 1:3311
2 3T 2 3T
0:7422 þ j0:06 0:7422  j0:06
6 0:2586  j27:632 7 6 0:2586 þ j27:632 7
6 7 6 7
6 0:000019 þ j0:00094 7 6 0:000019  j0:00094 7
6 7 6 7
6 0:0034 þ j0:0146 7 T 6 0:0034  j0:0146 7
 T6
w ¼6 7 w ¼ 6
6 0:0004 þ j0:000131 7 7 6 0:0004  j0:000131 7
7
6 7 6 7
6 0:000019 þ j0:000298 7 6 0:000019  j0:000298 7
6 7 6 7
4 0:000139 þ j0:0002727 5 4 0:000139  j0:0002727 5
0:0499 þ 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
wT8 ¼6
6 0:0283 7
7
6 7
6 0:000637 7
6 7
4 1:5496 5
0:079
78 2 A Single-Machine Infinite-Bus Power System …

According to Eq. (2.64), for the pair of electromechanical oscillation modes, the
residue is calculated to be

 6;7 ¼ w
R  T6;7 bo cTo v6;7
2 3T 2 3
0:7422  j0:06 0
6 0:2586
j27:632 7 6 0 7
6 7 6 7
6 7 6 7
6 0:000019  j0:00094 7 6 10000 7
6 7 6 7
6 0:0034  j0:0146 7 6 0 7
6 7 6 7
¼6 7 6 7
6 0:0004  j0:000131 7 6 0 7
6 7 6 7
6 0:000019  j0:000298 7 6 0 7
6 7 6 7
6 7 6 7
4 0:000139  j0:0002727 5 4 0 5
0:0499 þ j0:2027 0
2 3
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
 ½0 1 0 0 0 0 0 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
¼ 0:1772
j0:0114

From Eq. (2.104), the sensitivity of electromechanical oscillation modes to the


coefficient of damping torque provided by the PSS can be computed to be

6;7
@k  62;72 v62;72
w ð0:2586
j27:632Þð0:0016  j0:0188Þ
¼ ¼
@Dpss M 7
¼ 0:0743
j0:0056

At the complex frequency k6;7 ¼ 0:71  j8:44, the forward path can be cal-
culated to be
2.4 Examples 79

6;7 Þ ¼ aT ðsI  A33 Þ1 bpss3


Fpss ðk 238
¼ ½ 0 1:425 3:283 0:617 0:729 2:698 
0 2 3 11
1
B 6 7 C
B 6 1 7 C
B 6 7 C
B 6 1 7 C
B ð0:71  j8:44Þ6 7 C 2 3
B 6 7 C
B 6 1 7 C 10000
B 6 7 C 6
B 4 1 5 C 6 0 7 7
B C 6
B C 6 0 7 7
B 1 C
B 2 3C 6 7
B 100 6116:3 1404:6 2646 3127 1154:5 C 6 6 0 7
7
B C 6 7
B 6 0 6:65 7 C
135:03 7 C 4 0 5
B 6 478:4 2:88 3:40
B 6 7C
B 6 0 513:64 5:48 86:29 101:98 4:50 7 C 0
B 6 7C
B 6 314:16 1:22 7C
B 6 0:43 0 0:71 0 7C
B 6 7C
@ 4 0 1:36 0 1:88 3:49 0 5A
0 0 36:0 0 0 46:8
¼ 2:3823  j0:0253

Hence,

6;7
@k
R6;7 ¼  Fpss ðk6;7 Þ ¼ ð0:0743
j0:0056Þð2:3823  j0:0253Þ
@Dpss
¼ 0:1772
j0:0114

It is thus confirmed that the residue is equal to the forward path of the PSS
multiplied by the sensitivity of oscillation modes to the coefficient 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):316–329
2. Heffron WG, Phillips RA (1952) Effect of modern amplidyne voltage regulators on
underexcited operation of large turbine generators. AIEE Trans Power Apparatus Syst
71:692–697
3. Demello FP, Laskowski TF (1975) Concepts of power system dynamic stability. IEEE Trans
Power Apparatus Syst 94(3):827–833
4. Larsen EV, Swann DA (1981) Applying power system stabilizers part I: general concepts. IEEE
Trans Power Appar Syst 100(6):3017–3024
Chapter 3
Damping Torque Analysis
of Thyristor-Based FACTS Stabilizers
Installed in Single-Machine Infinite-Bus
Power Systems

3.1 A Single-Machine Infinite-Bus Power System Installed


with an SVC Stabilizer

3.1.1 Extended Heffron—Phillips Model of a Single-Machine


Infinite-Bus Power System Installed with an SVC
Stabilizer

3.1.1.1 Nonlinear Mathematical Model of a Single-Machine


Infinite-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
configuration of a thyristor-controlled reactor and fixed capacitor (TCR-FC) type of
SVC with an additional damping control loop, i.e., the SVC stabilizer.
Figure 3.2 shows the configuration of a single-machine infinite-bus power
system installed with an SVC, where bsvc is the equivalent admittance of the SVC.
For the TCR-FC, SVC [1]

jbsvc ¼ ½1  CðaÞ=jxsvcl  1=jxsvcc ð3:1Þ

where CðaÞ ¼ 2asin


2p
2a
 1 and a is the firing angle.
From Fig. 3.2, it can have

Isb ¼ Its  Is ¼ Its  jbsvc V


s
ð3:2Þ
V s ¼ jxsbIsb þ V b ¼ jxsbIts þ xsb bsvc V
s þ V
b

© Springer Science+Business Media New York 2016 81


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_3
82 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

x svcl
α0
Vs π

Vsref - x svcc
voltage α firing circuit
controller +
+

π/2

ysvc SVC
stabilizer

Fig. 3.1 Configuration of a TCP-FC SVC with damping control function

Fig. 3.2 A single-machine Vt Vs Vb


infinite-bus power system xts xsb
installed with an SVC

Its I sb

Is jb svc

 t ¼ jxtsIts þ V
V s ð3:3Þ

From Eq. (3.2), it can be obtained that

 
 s ¼ jxsb Its þ Vb
V ð3:4Þ
csvc

where csvc ¼ 1  xsb bsvc :


By substituting Eq. (3.4) into Eq. (3.3) it can have

 t ¼ jxlRIts þ V
V  b =csvc ð3:5Þ

where xlR ¼ xts þ 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 infinite bus with a voltage being cVsvcb .
3.1 A Single-Machine Infinite-Bus Power System Installed … 83

Therefore, the mathematical model of the single-machine infinite-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 modifi-
cation is to replace xt and Vb in Eqs. (2.37) and (2.38) by xlR and cVsvcb , respectively,
to obtain that
:
d ¼ xo ðx  1Þ
: 1
x ¼ ½Pm  Pt  D(x  1Þ
M
0 1 ð3:6Þ
E_ q ¼ 0 ðEq þ Efd Þ
Tdo
0 1 KA
E_ fd ¼  E0fd þ ðVtref  Vt Þ
TA TA

E0q Vb V2b ðxq  x0d Þ


Pt ¼ sin d  sin 2d
csvc x0dR c2svc 2 x0dR xqR
E0q xdR ðxd  x0d Þ Vb cos d
Eq ¼ 0 
xdR csvc x0dR ð3:7Þ
Efd ¼ Efd0 þ E0fd
xq Vb sin d xt E0q Vb x0d cos d qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
vtd ¼ ; vtq ¼ 0 þ 0 ; Vt ¼ v2td + v2tq
csvc xqR xdR csvc xdR

where x0dR ¼ x0d þ xlR ; xdR ¼ xd þ xlR ; xqR ¼ xq þ xlR :


Similar modification of Eq. (2.33) gives

E0q Vb cos d
itsd ¼ 
x0dR csvc x0dR
ð3:8Þ
Vb sin d
itsq ¼
csvc xqR

where itsd and itsq are the d and q component of line current Its , respectively.

3.1.1.2 Extended Heffron–Phillips Model

By linearizing the first two equations in Eq. (3.7), it can have

DPt ¼ K1 Dd þ K2 DE0q þ Kp Dbsvc


ð3:9Þ
DEq ¼ K3 DE0q þ K4 Dd þ Kq Dbsvc
84 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

where

E0q0 Vb V2b ðxq  x0d Þ


K1 ¼ cos d0  cos 2d0
csvc0 x0dR c2svc0 x0dR xqR
Vb
K2 ¼ sin d0
csvc0 x0dR
 0
   
@Pt  Vb Eq0 sin d0 xsb xts þ x0d V2b ðxq  x0d Þsin(2d0 Þxsb xts þ x0d
Kp ¼ ¼ 
@bsvc 0 ðcsvc0 x0dR Þ2 2c3svc0 ðx0dR Þ2 xqR
 
V2 ðxq  x0d Þsin(2d0 Þxsb xts þ xq
 b
2c3svc0 x0dR ðxqR Þ2
  
xsb xts þ x0d Vb E0q0 sin d0 V2b ðxq  x0d Þsin(2d0 Þ
¼ 
csvc0 x0dR csvc0 x0dR 2c2svc0 x0dR xqR
 
Vb ðxq  x0d Þsin(2d0 Þ xsb xts þ xq
2

2c2svc0 x0dR xqR csvc0 xqR
xdR
K3 ¼
x0dR
ðxd  x0d Þ Vb sin d0
K4 ¼
csvc0 x0dR
E0q0 xdR x2 E0q0 x2 ðxd  x0d Þ Vb cos d0  
Kq ¼  0 2 2sb þ 0 2sb  0 2
xsb xts þ x0d
ðxdR Þ csvc0 x dR csvc0 ðcsvc0 xdR Þ

In deriving Kp and Kq above, following equations are used

csvc x0dR ¼ csvc ðx0d þ xts þ xsb =csvc Þ ¼ csvc ðx0d þ xts Þ þ xsb
¼ ð1  xsb bsvc Þðx0d þ xts Þ þ xsb ð3:10Þ
csvc xq ¼ ð1  xsb bsvc Þðxq þ xts Þ þ xsb

By linearizing the last three equations in Eq. (3.7), it can have

xq Vb cos d0 xq Vb sin d0
Dvtd ¼ Dd þ xsb ðxsb þ xq ÞDbsvc
csvc0 xqR ðcsvc0 xqR Þ2
Vb x0d sin d0 xLR
Dvtq ¼  Dd þ 0 DE0q ð3:11Þ
csvc0 x0dR xdR
" #
xlR E0q0 x2 E0q0 x2 Vb x0d cos d0
þ  0 2 2sb þ 0 2sb þ xsb ðxsb þ x0d Þ Dbsvc
ðxdR Þ csvc0 xdR csvc0 ðcsvc0 x0dR Þ2
3.1 A Single-Machine Infinite-Bus Power System Installed … 85

Hence,
vtd0 vtq0
DVt ¼ Dvtd þ Dvtq ¼ K5 Dd þ K6 DE0q þ Kv Dbsvc ð3:12Þ
Vt0 Vt0

where

vtd0 xq Vb cos d0 vtq0 Vb x0d sin d0


K5 ¼ 
Vt0 csvc0 xqR Vt0 csvc0 x0dR
vtq0 xLR
K6 ¼
Vt0 x0dR
vtd0 xq Vb sin d0
Kv ¼ xsb ðxsb þ xq Þ
Vt0 ðcsvc0 xqR Þ2
" #
vtq0 xlR E0q0 x2sb E0q0 x2sb Vb x0d cos d0 0
þ  0 2 2 þ 0 2 þ x ðx þ xd Þ
0 Þ2 sb sb
Vt0 ðxdR Þ csvc0 xdR csvc0 ðcsvc0 xdR

By substituting Eqs. (3.9) and (3.12) into the linearized equation of Eq. (3.6),
it can have

Dd_ ¼ xo Dx
1
Dx_ ¼ ðK1 Dd  DDx  K2 DE0q  Kp Dbsvc Þ
M
_ 0 1 ð3:13Þ
DEq ¼ 0 ðK3 DE0q  K4 Dd  Kq Dbsvc þ DEfd Þ
Td0
0 1 0 KA
DE_ fd ¼  DE_ fd  ðK5 Dd þ K6 DE0q þ Kv Dbsvc Þ
TA TA

Equation (3.13) is the extended Heffron–Phillips model of the single-machine


infinite-bus power system installed with the SVC, which is shown in Fig. 3.3.

3.1.1.3 Extended Heffron–Phillips Model with Both the SVC Voltage


and Damping Control Function Included

By substituting Its ¼ itsd þ jitsq into Eq. (3.4), it can have

Vb sin d  itsq xsb


vsd ¼
csvc
ð3:14Þ
Vb cos d þ itsd xsb
vsq ¼
csvc
86 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

K1

1 ω0
- D + sM s

Δbsvc
K2 K4 K5
Kp Kq Kv
SVC

1
- - KA
K 3 + sTd0 ' + 1 + sTA
-

K6

Fig. 3.3 Extended Heffron–Phillips model of the single-machine infinite-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

Vb sin d  V b sin d 
csvc xqR xsb 1 xsb Vb sin d
vsd ¼ ¼ Vb sin d 
csvc csvc csvc xqR
 0
1 xsb Eq Vb cos d ð3:15Þ
vsq ¼ Vb cos d þ 0  xsb
csvc xdR csvc x0dR
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Vs ¼ v2sd þ v2sq

Linearization of Eq. (3.15) is as follows:

DVs ¼ C1 Dd þ C2 DE0q þ C3 Dbsvc ð3:16Þ

where
  
@Vs  vsd0 @vsd  vsq0 @vsq 
C1 ¼ ¼ þ
@d 0 Vs0 @d 0 Vs0 @d 0
 
vsd0 1 xsb Vb cos d0 vsq0 1 Vb sin d0
¼ Vb0 cos d0  þ Vb0 sin d0 þ x sb
Vs0 csvc0 csvc0 xqR Vs0 csvc0 csvc0 x0dR
3.1 A Single-Machine Infinite-Bus Power System Installed … 87

  
@Vs  vsd0 @vsd  vsq0 @vsq  vsq0 1 xsb
C2 ¼ 0  ¼  þ  ¼
@Eq  Vs0 @E0q  Vs0 @E0q  Vs0 csvc0 x0dR

0 0
 0

@Vs  vsd0 @vsd  vsq0 @vsq 
C3 ¼ ¼ þ
@bsvc 0 Vs0 @bsvc 0 Vs0 @bsvc 0
" #
vsd0 xsb vsd0 x2sb Vb sin d0
¼  ðxq þ xts Þ
Vs0 csvc0 csvc0 ðcsvc0 xqR Þ2


vsq0 xsb xsb E0q0 Vb cos d0
þ V b cos d 0 þ  x
x0dR csvc0 x0dR
sb
Vs0 c2svc0
 
1 E0q0 x3 Vb cos d0
þ  0 2 2sb  2 0 2 x2sb ðx0d þ xts Þ
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 ¼ CV DVs þ CS Dx ð3:17Þ

From Eq. (3.1), it can be obtained that

1 1  C(aÞ
bsvc ¼  ð3:18Þ
xsvcc xsvcl

Hence,

_ 0Þ
Cða _ 0Þ
Cða
Dbsvc ¼ Da ¼ ðCV DVs þ CS DxÞ
xsvcl xsvcl ð3:19Þ
where
4 _ 0 Þ ¼ ð2  2cos 2a0 Þ [ 0
[ Cða
p p

By substituting Eq. (3.16) into Eq. (3.19), the following equation can be
obtained
:
Cða0 Þ
Dbsvc ¼ ðCV C1 Dd  CV C2 DE0q  CV C3 Dbsvc þ CS DxÞ ð3:20Þ
Xsvcl

Thus,

CV C1 Dd  CV C2 DE0q þ CS Dx
Dbsvc ¼ x ð3:21Þ
: svcl þ CV C3
Cða0 Þ
88 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

K1

1 ω0
- D + sM s

SVC Cs
K2 C + C3 C V

−C 2 C V −C1CV
+
C + C3 C V C + C3 C V

Δbsvc
K4 K5

Kp Kq Kv

1
- - KA
K 3 + sTd0 ' 1 + sTA
-
+

K6

Fig. 3.4 Extended Philip–Heffron 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 ¼ :xsvcl .
Cða0 Þ

3.1.1.4 Calculation of Initial Compensation

This section demonstrates how the initial compensation of the SVC installed in the
single-machine infinite-bus power system shown in Fig. 3.2 can be calculated to
3.1 A Single-Machine Infinite-Bus Power System Installed … 89

satisfy the requirement to maintain the voltage profile 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
V b 
 ¼ Psb0  jQsb0
V ð3:22Þ
s0
jxsb

Since the resistance of the transmission lines is ignored, Psb0 ¼ Pt0 . By choosing
 s0 as the reference phasor on x-axis, it can have V
V   ¼ Vs0 . From
 s0 ¼ V
s0
Eq. (3.14), the following equation can be obtained

 b ¼ Vs0  xsb ðQsb0 þ jPsb0 Þ ¼ Vs0  xsb Qsb0  j xsb Psb0


V
Vs0 Vs0 Vb

The above equation gives


 2  2
xsb xsb
V2b ¼ Vs0  Q þ Psb0 ð3:23Þ
Vs0 sb0 Vs0

Solution of Eq. (3.23) is as follows


2 sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3
 2ffi
Vs0 4 x
Psb0 5
sb
Qsb0 ¼ Vs0  V2b  ð3:24Þ
xsb Vs0

Pt 0 + jQ t 0 jx ts Pts0 + jQ ts0 Psb0 + jQsb0 jx sb Pb0 + jQ b0


Vt0 V s0 Vb

I ts0 Isb0
Ps0 + jQs0

bsvc0
j Is

Fig. 3.5 Circuit model of the single-machine infinite-bus power system installed with SVC
90 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

Similarly, from the following equation

 t0  V
V  s0 
 ¼ Pts0  jQts0
V s0
jxts

it can have
 2  2
xts xts
V2t0 ¼ Vs0 þ Q þ Pts0 ð3:25Þ
Vs0 ts0 Vs0

Solution of Eq. (3.25) is as follows:


2sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3
 2
Vs0 4 x
Pts0  Vs0 5
ts
Qts0 ¼ V2t0  ð3:26Þ
xts Vs0

Hence,

Ps0 þ jQs0 ¼ 0:0 þ j(Qts0  Qsb0 Þ

Since

 s0 jbsvc0 ¼ Is0
V

it can have

  jbsvc0 ¼ Is0 V
 s0 V
V   ¼ Ps0  jQs0
s0 s0

Finally from the above equation, the initial compensation is calculated to be

ðPs0  jQs0 Þ Q Q Q
bsvc0 ¼ ¼  s0 ¼ sb0 2 ts0 ð3:27Þ
jV2s0 V2s0 Vs0

3.1.2 Damping Torque Analysis of SVC Stabilizer

3.1.2.1 Electric Torque Provided by the SVC Stabilizer

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
3.1 A Single-Machine Infinite-Bus Power System Installed … 91

Δω
-

Cs
K2 C + C3 C V

−C 2 C V
+
C + C3 C V

Δbsvc

Kp Kq Kv

1
- KA
K 3 + sTd0 ' 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 ¼ DTsd þ DTdd ¼ Tsd Dd þ Tdd Dx


ð3:28Þ
DTei ¼ DTsi þ DTdi ¼ Tsi Dd þ Tdi Dx

where Tsd ; Tsi ; Tdd ; Tdi are direct, indirect synchronizing torque and direct, indirect
damping torque coefficients, 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 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

1 1 KA
K3 þ sT0d0 K3 þ sT0d0 1 þ sTA
Fsvc ðs) ¼ Kp  K2 Kq  K2 Kv
1 þ K6 1 KA 1 þ K6 1 KA
K3 þ sT0d0 1 þ sTA K3 þ sT0d0 1 þ sTA
K2 ½Kq ð1 þ sTA Þ þ Kv KA 
¼ Kp 
ðK3 þ sT0d0 Þð1 + sTA Þ + K6 KA
ð3:29Þ

Hence, the electric torque provided by the SVC stabilizer is as follows

DTet ¼ Fsvc ðjxs ÞDbsvc ð3:30Þ

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 coefficient Kp is the
dominant weight parameter in determining the amount of damping torque contri-
bution from the SVC stabilizer. Hence, in the following section, how the effec-
tiveness 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

From Eq. (3.7), it can have

E0q0 Vb V2b ðxq  x0d Þ


Pt0 ¼ sin d0  sin 2d0 ¼ Pt10  Pt20 ð3:31Þ
csvc0 x0dR c2svc0 2 x0dR xqR

E0q0 Vb V2b ðxq x0d Þ


where Pt10 ¼ csvc0 x0dR sin d0 ; Pt20 ¼ 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 þ x0d Vb E0q0 sin d0 V2b ðxq  x0d Þsin(2d0 Þ
Kp ¼ 
csvc0 x0dR csvc0 x0dR 2c2svc0 x0dR xqR
0
 
Vb ðxq  xd Þsin(2d0 Þ xsb xts þ xq
2
 ð3:32Þ
2c2svc0 x0dR xqR csvc0 xqR
 0
   
xsb xts þ xd xts þ xq
¼ Pt0  Pt20
csvc0 x0dR xqR

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.
3.1 A Single-Machine Infinite-Bus Power System Installed … 93

Fig. 3.7 Pt  d curve of the Pt 0 = Pt10 − Pt 20


power system

Pt10

Pt 20

π/2 δ0
π/4 π

3.1.2.3 Influence of Parameters of the Generator

Since xq [ x0d , it should have


   
xts þ x0d 1 1 xts þ xq
¼ \ ¼ ð3:33Þ
x0dR 1 þ c xxsbþ x0 1 þ c xxsbþ x xqR
svc0 ð ts dÞ svc0 ð ts qÞ

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 0.2–0.35 and that of xq
is around 0.45–0.7) or if the system operates at a higher load condition
ðxts þ x0 Þ ðxts þ xq Þ
(Pt0  Pt20 ), x0 d Pt0 [ xqR Pt20 such that Kp [ 0, the SVC stabilizer
dR
will provide the power system with positive damping.
ðxts þ x0 Þ ðxts þ xq Þ
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 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

1.2–2.2) 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
ðxts þ x0d Þ ðxts þ xq Þ
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.
ðxts þ x0 Þ ðxts þ xq Þ
3. Around the operating point where x0 d Pt0 ¼ xqR Pt20 such that Kp ¼ 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 stabilizer’s
function.

3.1.2.4 Electric Length of the Transmission Line

Define z to be an index of the length of the transmission line and take z0 ¼ 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))
     
zxsb zxts þ x0d zxts þ xq
Kp ðz) ¼ Pt0  Pt20 ð3:34Þ
csvc0 x0d þ z(xts þ xsb =csvc0 Þ xq þ z(xts þ xsb =csvc0 Þ

For the simplicity of discussion, it is assumed that at the steady-state operation,


bsvc0 ¼ 0 such that csvc0 ¼ 1. Hence,
     
zxts þ x0d zxts þ xq
Kp ðz) ¼ zxsb Pt0  Pt20 ð3:34Þ
x0d þ z(xts þ xsb Þ xq þ z(xts þ xsb Þ

It can have
     
@ zxts þ x0d zxts þ xq
Kp ðz) ¼ xsb 0 Pt0  Pt20
@z xd þ z(xts þ xsb Þ xq þ z(xts þ xsb Þ
(  
xts zxts þ x0d ðxts þ xsb Þ
þ zxsb 0 Pt0  Pt0
xd þ z(xts þ xsb Þ ½x0d þ z(xts þ xsb Þ2
  )
xts zxts þ xq ðxts þ xsb Þ
 Pt20 þ Pt20
xq þ zðxts þ xsb Þ ½xq þ z(xts þ xsb Þ2
3.1 A Single-Machine Infinite-Bus Power System Installed … 95

Hence,
      
@Kp ðz) xts þ x0d xts þ xq
¼ x P  P
@z z¼z0 ¼1
sb to t02
x0d þ ðxts þ xsb Þ xq þ ðxts þ xsb Þ
(  
xts xts þ x0d ðxts þ xsb Þ
þ xsb 0 Pt0  0 Pt20
xd þ ðxts þ xsb Þ ½xd þ ðxts þ xsb Þ2
  )
xts xts þ xq ðxts þ xsb Þ
 Pt20 þ Pt20
xq þ ðxts þ xsb Þ ½xq þ ðxts þ xsb Þ2
   
xsb xts þ x0d xsb xts xsb xts þ x0d ðxts þ xsb Þ
¼ 0 Pt0 þ 0 Pt0  Pt0
xd þ ðxts þ xsb Þ xd þ ðxts þ xsb Þ ðx0d þ xts þ xsb Þ2
   
xsb xts þ xq xsb xts xsb xts þ xq ðxts þ xsb Þ
 Pt20  Pt20 þ Pt20
xq þ ðxts þ xsb Þ xq þ ðxts þ xsb Þ ðxq þ xts þ xsb Þ2
   
xsb xts þ x0d xsb xts xsb xts þ x0d ðxts þ xsb Þ
¼ 0 Pt0 þ 0 Pt0  Pt0
xd þ ðxts þ xsb Þ xd þ ðxts þ xsb Þ ðx0d þ xts þ xsb Þ2
   
xsb xts þ xq xsb xts xsb xts þ xq ðxts þ xsb Þ
 Pt20  Pt20 þ Pt20
xq þ ðxts þ xsb Þ xq þ ðxts þ xsb Þ ðxq þ xts þ xsb Þ2
   
xsb xts þ x0d x0d þ xsb xts x0d þ xts þ xsb
¼ Pt0
ðx0d þ xts þ xsb Þ2
   
xsb xts þ xq xq þ xsb xts xq þ xts þ xsb
  2 Pt20
xq þ xts þ xsb
   
xsb x0 xts þ x0d xsb xts xsb xq xts þ xq xsb xts
¼ 0 d P t0 þ 0 P t0    P 
2 t20
Pt20
ðxd þ xts þ xsb Þ2 xd þ xts þ xsb xq þ xts þ xsb x q þ xts þ xsb
   
xsb x0d xts þ x0d xsb xq xts þ xq xsb xts xsb xts
¼ Pt0  Pt20 þ 0 Pt0  Pt20
x02
dR x2qR xdR xqR
    
xsb x0d xts þ x0d xsb xq xts þ xq Pt0 Pt20
¼ 02
Pt0  2
Pt20 þ xsb xts 0 
xdR xqR xdR xqR
ð3:35Þ

Thus,
    
@Kp ðz) xsb x0d xts þ x0d þ xsb xts x0dR xsb xq xts þ xq þ xsb xts xqR
¼ P  Pt20
@z z¼z0 ¼1
t0
X02
dR X2qR
 2  2
xsb xts þ x0d þ xts x2sb xsb xts þ xq þ xts x2sb
¼ Pt0  Pt20
x02
dR x2qR
ð3:36Þ

When Kp [ 0
   
xts þ x0d xts þ xq
Pt0 [ Pt20 ð3:37Þ
x0dR xqR
96 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

Hence, normally it should have


 2  2
xts þ x0d xts þ xq
Pt0 [ Pt20 ð3:38Þ
x0dR xqR

Using Eqs. (3.36) and (3.38), from Eq. (3.35), it can be obtained that
@Kp ðz)
[0 ð3:39Þ
@z

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 fixed, that is,
xts  xsb ¼ xt , from Eq. (3.32), it can be obtained that (assuming csvc0 ¼ 1)
     
xsb xt  xsb þ x0d xsb xt  xsb þ xq
Kp ðxsb Þ ¼ P  P ð3:40Þ
xt þ x0d
t0 t20
xq þ xt

Define the following two functions of xsb


 
xsb xt  xsb þ x0d
g(xsb Þ ¼ Pt0
x0 þ xt
 d  ð3:41Þ
xsb xt  xsb þ xq
h(xsb Þ ¼ Pt20
xq þ xt

Thus,

Kp ðxsb Þ ¼ g(xsb Þ  h(xsb Þ ð3:42Þ

By solving the following two equations

@ xt  2xsb þ x0d
g(xsb Þ ¼ Pt0 ¼ 0
@xsb x0d þ xt
ð3:43Þ
@ xt  2xsb þ xq
h(xsb Þ ¼ Pt20 ¼ 0
@xsb xq þ xt

ðx0d þ xt Þ ðx þ x Þ
It can have xsb ¼ ¼ xsb1 and xsb ¼ q 2 t ¼ xsb2 . Hence, g(xsb Þ and
2
0
ðx þ x Þ ðx þ x Þ
h(xsb Þ achieve its maximum value at xsb ¼ d 2 t ¼ xsb1 and xsb ¼ q 2 t ¼ xsb2 ,
respectively.
3.1 A Single-Machine Infinite-Bus Power System Installed … 97

K p (x sb )

g(x sb )

K p (x sb )

h(x sb )

x sb
0 xt x sb1 x sb2 xt
2

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 ðxsb Þ can be sketched as shown in
Fig. 3.8 (Pt0  Pt20 ). Figure 3.8 demonstrates that when xsb  x2t , Kp ðxsb Þ 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 A Single-Machine Infinite-Bus Power System Installed


with a TCSC or TCPS Stabilizer

3.2.1 Extended Heffron–Phillips Model of a Single-Machine


Infinite-Bus Power System Installed with a TCSC
or TCPS Stabilizer

3.2.1.1 Extended Heffron–Phillips Model of a Single-Machine


Infinite-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 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

(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 infinite-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 infinite busbar. Therefore, the
mathematical model of the system can be established by simply modifying Eqs. (2.
37) and (2.38). The modification is just to change the expressions of the equivalent
reactance of the transmission line as follows

d_ ¼ xo ðx  1Þ
1
x_ ¼ ½Pm  Pt  D(x  1Þ
M
0 1 ð3:44Þ
E_ q ¼ 0 ðEq þ Efd Þ
Tdo
0 1 KA
E_ fd ¼  E0fd þ ðVtref  Vt Þ
TA TA
 
E0q Vb V2b xq  x0d
Pt ¼ 0 sin d  sin 2d
xdR 2 x0dR xqR
 
E0q xdR xd  x0d Vb cos d
Eq ¼ 0 
xdR x0dR ð3:45Þ
Efd ¼ Efd0 þ E0fd
xq Vb sin d xt E0q Vb x0d cos d qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
vtd ¼ ; vtq ¼ 0 þ ; Vt ¼ v2td þ v2tq
xqR xdR x0dR

where

xdR ¼ xd þ xL  xtcsc ; x0dR ¼ x0d þ xt  xtcsc


ð3:46Þ
xqR ¼ xq þ xL  xtcsc

Vt Vb
− jx tcsc

jx t

Fig. 3.9 A single-machine infinite-bus power system installed with a TCSC


3.2 A Single-Machine Infinite-Bus Power System Installed … 99

By linearizing Eq. (3.45), it can be obtained that

DPt ¼ K1 Dd þ K2 DE0q þ Kp Dxtcsc


DEq ¼ K3 DE0q þ K4 Dd þ Kq Dxtcsc ð3:47Þ
DVt ¼ K5 Dd þ K6 DE0q þ Kv Dxtcsc

where
 
E0q0 Vb V2b xq  x0d
K1 ¼ cos d0  cos 2d0
x0dR x0dR xqR
Vb
K2 ¼ 0 sin d0
xdR
 0
@Pt  Vb Eq0 sin d0 V2b ðxq  x0d Þ sin 2d0  0 
Kp ¼  ¼ 02
 02
xdR þ xqR
@x tcsc 0 x dR 2x x 2
dR qR
xdR
K3 ¼ 0
xdR
 
xd  x0d Vb sin d0
K4 ¼
x0dR
 0  
@Eq  Eq0 xdR xd  x0d Vb cos d0
Kq ¼ ¼ 02 
@xtcsc 0 xdR x02
dR
0
vtd0 xq Vb cos d0 vtq0 Vb xd sin d0
K5 ¼ 
Vt0 xqR Vt0 x0dR
vtq0 xLR
K6 ¼
Vt0 x0dR
  
@Vt  vtd0 xq Vb sin d0 vtq0 xL 1 0 Vb x0d cos d0
Kv ¼ ¼ þ  0 Eq0 þ
@xtcsc 0 Vt0 x2qR Vt0 x02
dR xdR x02dR

By substituting Eq. (3.46) into the linearized equation of Eq. (3.44), it can have
:
D d ¼ xo Dx
: 1
D x ¼ ðK1 Dd  DDx  K2 DE0q  Kp Dxtcsc Þ
M
0 1 ð3:48Þ
DE_ q ¼ 0 ðK3 DE0q  K4 Dd  Kq Dxtcsc þ DE0fd Þ
Td0
1 KA
DE0fd ¼  DE0fd  ðK5 Dd þ K6 DE0q þ Kv Dxtcsc Þ
TA TA
100 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

K1

1 ω0
- D + sM s

Δx t csc
K2 K4 K5
Kp Kq Kv
TCSC

1
- - KA
K 3 + sTd0 ' + 1 + sTA
-

K6

Fig. 3.10 Extended Heffron–Phillips model of the single-machine infinite-bus power system
installed with a TCSC stabilizer

Equation (3.48) is the extended Heffron–Phillips model of the single-machine


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

3.2.1.2 Extended Heffron–Phillips Model of a Single-Machine


Infinite-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 dec-
ades 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 infinite-bus power system installed
with a TCPS as shown in Fig. 3.12, system dynamic Eqs. (2.37) and (2.38) need to
be modified only in the phase relationship as follows.
3.2 A Single-Machine Infinite-Bus Power System Installed … 101

a V V'

Vector diagram
Symbol
V V'
V
φ
V'
Converter

TCPS

Fig. 3.11 Arrangement of a TCPS and its simple representation

Vt ∠δ Vt ∠δ + φ Vb

TCPS jx t

Fig. 3.12 The single-machine infinite-bus power system installed with a TCPS

:
d ¼ xo ðx  1Þ
: 1
x ¼ ½Pm  Pt  D(x  1Þ
M
ð3:49Þ
_E0 ¼ 1 ðEq þ Efd Þ
q
T0do
0 1 KA
E_ fd ¼  E0fd þ ðVtref  Vt Þ
TA TA

E0q Vb V2b ðxq  x0d Þ


Pt ¼ sin(d þ /Þ  sin2(d þ /Þ
x0dR 2 x0dR xqR
E0q xdR ðxd  x0d ÞVb cos(d þ /Þ
Eq ¼ 0 
xdR x0dR
ð3:50Þ
xq Vb sin(d þ /Þ
vtd ¼
xqR
xL E0q Vb x0d cos(d þ /Þ
vtq ¼ 0 þ
xdR x0dR
102 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

By linearizing Eq. (3.50), it can have

DPt ¼ K1 Dd þ K2 DE0q þ Kp D/
DEq ¼ K3 DE0q þ K4 Dd þ Kq D/ ð3:51Þ
DVt ¼ K5 Dd þ K6 DE0q þ Kv D/

where

E0q0 Vb V2b ðxq  x0d Þ


K1 ¼ cos d0  cos 2d0
x0dR x0dR xqR
Vb
K2 ¼ 0 sin d0
xdR
 0
@Pt  Eq0 Vb V2 ðxq  x0d Þ
Kp ¼  ¼ 0 cos d0  b 0 cos 2d0
@/ 0 x dR x x dR qR
xdR
K3 ¼ 0
xdR
ðxd  x0d ÞVb sin d0
K4 ¼
x0dR

@Eq  ðxd  x0d ÞVb sin d0
Kq ¼ ¼
@/ 0 x0dR
vtd0 xq Vb cos d0 vtq0 Vb x0d sin d0
K5 ¼ 
Vt0 xqR Vt0 x0dR
vtq0 xLR
K6 ¼
Vt0 x0dR

@Vt  vtd0 xq Vb cos d0 vtq0 Vb x0d sin d0
Kv ¼ ¼ 
@/ 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 ¼ xo Dx
: 1
D x ¼ ðK1 Dd  DDx  K2 DE0q  Kp D/Þ
M
0 1 ð3:52Þ
DE_ q ¼ 0 ðK3 DE0q  K4 Dd  Kq D/Þ
Td0
1 KA
DE0fd ¼  DE0fd  ðK5 Dd þ K6 DE0q þ Kv D/Þ
TA TA

Equation (3.52) is the extended Heffron–Phillips model of the single-machine


infinite-bus power system installed with the TCSC stabilizer, which is shown in
Fig. 3.13.
3.2 A Single-Machine Infinite-Bus Power System Installed … 103

K1

1 ω0
- D + sM s

Δφ
K2 K4 K5
Kp Kq Kv
TCPS

1
- - KA
K 3 + sTd0 ' + 1 + sTA
-

K6

Fig. 3.13 Extended Heffron–Phillips model of the single-machine infinite-bus power system
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 ðs). The output control signal
of the TCSC and TCPS stabilizers is as follows

Dutc ¼ Ktc Ttc ðs)Dx ð3:53Þ

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 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

ΔTd

-
ΔTdd
ΔTdi
Δu tc
K2

Kp Kq Kv

1
- KA
K 3 + sTd0 ' + 1 + sTA
-

K6

Fig. 3.14 Forward path of the TCSC or TCPS stabilizer

because Dutc is attenuated by two first-order lag blocks before it forms the indirect
damping torque. That is

DTd  Tdd Dx ¼ Re[Kp Ktc Ttc ðjxs ÞDx ð3:54Þ

From Eq. (3.54), it can be seen that coefficient 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]

From Eq. (3.45), it can have


 
E0q0 Vb V2b xq  x0d
Pt0 ¼ 0 sin d0  sin 2d0 ¼ Pt10  Pt20 ð3:55Þ
xdR 2 x0dR xqR

where

E0q0 Vb
Pt10 ¼ sin d0
x0dR
ð3:56Þ
V2 ðxq  x0d Þ
Pt20 ¼ b 0 sin 2d0
2 xdR xqR
3.2 A Single-Machine Infinite-Bus Power System Installed … 105

Hence, for the TCSC stabilizer, from Eq. (3.47), it can be obtained that
 0
@Pt  Vb Eq0 sin d0 V2b ðxq  x0d Þ sinð2d0 Þ  0 
Kp ¼  ¼ 02
 02
xdR þ xqR
@xtcsc 0 xdR 2
2xdR XqR
 0   
1 Vb Eq0 sin d0 Vb ðxq  xd Þsin(2d0 Þ
2 0
1 V2b xq  x0d sinð2d0 Þ
¼ 0  
xdR x0dR 2x0dR xqR xqR 2x0dR xqR
1 1
¼ 0 Pt10  Pt20
xdR xqR
ð3:57Þ

Since xqR [ x0dR and Pt0 [ Pt20 as shown by Fig. 3.15,

Kp [ 0 ð3:58Þ

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 Pt 0 = Pt10 − Pt 20


power system installed with a
TCSC
Pt10

Pt 20

π/2 δ0
π/4 π
106 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

effectiveness of the TCSC stabilizer in damping power system oscillations increases


with the increase in system load conditions.
From Eq. (3.57), it can have
!
@Kp Pt0 Pt20
¼  02  2 \0 ð3:59Þ
@xt xdR xqR

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 ¼ ¼ Kp1  Kp2 ð3:60Þ
@/ 0

where

Vb E0q0 cos d0
Kp1 ¼
x0
 dR 0  ð3:61Þ
V xq  xd cos 2d0
2
Kp2 ¼ b
2x0dR xqR

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.
From Eqs. (3.60) and (3.61), it can have
!
@Kp Kp1 Kp2
¼  02  2 \0 ð3:62Þ
@xt xdR xqR

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 b

π/2 δ0
π/4 π

Vb E q0 'sin δ0 Vb 2 (x q − x d ') sin 2 δ0


a= ,
b=
x dΣ ' 2x dΣ ' x qΣ

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 infinite-bus power system with an SVC


installed in Fig. 3.2 are (in p.u. except indicated)
0
Generator: xd ¼ 1:0; xq ¼ 0:8; x0d ¼ 0:15; M ¼ 6:0 s:; D ¼ 0; Td0 ¼ 5:044 s:
The AVR: KA ¼ 20:0; TA ¼ 0:01 s:
Transmission line: Xts ¼ 0:3; Xsb ¼ 0:3:
The SVC: xsvcl ¼ 1; xsvcc ¼ 1; Kvp ¼ 1; Kvi ¼ 8:
Steady-state operating condition: Pt0 ¼ 0:5; Vt0 ¼ 1:0; Vs0 ¼ 1:0; Vb ¼ 1:0
108 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

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 proportional–integral (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 ¼ a0 þ Kvp þ ðVsref  Vs Þ þ usvcs ð3:63Þ
s

where usvc-s is the output stabilizing signal of the SVC stabilizer, i.e.,

ð1 þ sT2 Þð1 þ sT4 Þ


usvcs ¼ Ks ðPtref  Pt Þ
ð1 þ sT1 Þð1 þ sT3 Þ

Linearization of Eq. (3.63) is as follows



Kvi
Da ¼  Kvp þ DVs þ Dusvcs ð3:64Þ
s
:

From Eq. (3.19) with Dbsvc ¼ Cxða 0Þ


svcl
Da, the above equation can be converted to
:  :
Cða0 Þ Kvi Cða0 Þ
Dbsvc ¼ Kvp þ DVs þ Dusvcs ð3:65Þ
xsvcl s xsvcl

Substituting Eq. (3.16) into the above equation, it can have


:   i C: ða Þ
Cða0 Þ Kvi h
C1 Dd þ C2 DE0q þ C3 Dbsvc þ
0
Dbsvc ¼  Kvp þ Dusvcs
xsvcl s xsvcl
ð3:66Þ

The above equation can give

Dbsvc ¼ Fsvc1 ðs)(C1 Dd þ C2 DE0q Þ þ Fsvc2 ðs)Dusvcs ð3:67Þ

where
:
:
 Cxða 0Þ
K vp þ Kvi
s 1 Cða0 Þ
Fsvc1 ðs) ¼ ; Fsvc2 ðs) ¼
svcl
: :
1 þ C3 Cxða 0Þ
K þ Kvi 1 þ C Cða0 Þ K þ Kvi xsvcl
svcl
vp s 3 xsvcl vp s
3.3 An Example Power System Installed with an SVC Stabilizer 109

K1

1 ω0
-
D + sM s

SVC
Fsvc2 (s)
K2 Δu svc −s

C2 Fsvc1 (s) + C1Fsvc1 (s)

Δbsvc
K4 K5

Kp Kq Kv

1
- - KA
K 3 + sTd0 ' 1 + sTA
-
+

K6

Fig. 3.17 Linearized model with SVC voltage control function included

From Eq. (3.67) and Fig. 3.3, linearized Heffron–Phillips 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.
:

In Eq. (3.66), denote Dzsvc ¼  Cxða 0 Þ Kvi


svcl
0
s ðC1 Dd þ C2 DEq þ C3 Dbsvc Þ, that is
:
: Cða0 Þ
D z ¼ Kvi ðC1 Dd þ C2 DE0q þ C3 Dbsvc Þ ð3:68Þ
svc xsvcl

Equation (3.66) becomes


:
_ 0Þ
Cða Cða0 Þ
Dbsvc ¼ Kvp C1 Dd þ C2 DE0q þ C3 Dbsvc þ Dz þ Dusvcs ð3:69Þ
xsvcl xsvcl
110 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

The above equation can give


_
 Cxða 0Þ
Dd þ DE 0 C_ ða0 Þ
K vp C1 C 2 q þ Dz þ xsvcl Dusvcs
Dbsvc ¼ ð3:70Þ
svcl
_
1 þ Cxða 0Þ
svcl
Kvp C3

Substituting Eq. (3.70) into Eq. (3.68), it can have

D_zsvc ¼ CA Dd þ CE DE0q þ CZ Dz þ CU Dusvcs ð3:71Þ

where
0: 1 :
: Cða0 Þ K C Cða0 Þ K C
Cða0 Þ B xsvcl vp 3 C xsvcl vi 3
CA ¼  Kvi C1 @1  : A; CZ ¼  : ;
xsvcl
1 þ Cxða 0Þ
svcl
K C
vp 3 1 þ Cxða 0Þ
svcl
K C
vp 3

0 1  : 2
:
: Cða0 Þ K C Cða0 Þ K C
Cða0 Þ B xsvcl vp 3 C xsvcl vi 3
CE ¼  Kvi C2 @1  : A; CU ¼  :
xsvcl
1 þ Cxða 0Þ
svcl
K C
vp 3 1 þ Cxða 0Þ
svcl
Kvp C3

Substituting Eq. (3.70) into Eq. (3.13), it can be obtained that


:
D d ¼ xo Dx
: 1
D x ¼ ðK01 Dd  DDx  K02 DE0q  Kpz Dz  K0p Dusvcs Þ
M
0 1 ð3:72Þ
DE_ q ¼ 0 ðK03 DE0q  K04 Dd  Kqz Dz  K0q Dusvcs þ DEfd Þ
Td0
0 1 0 KA 0
DE_ fd ¼  DE_ fd  ðK Dd þ K06 DE0q þ Kvz Dz þ K0v Dusvcs Þ
TA TA 5

where
: :

0
 Cxða 0Þ
Kvp C1 0
 Cxða 0Þ
Kvp C2
K1 ¼ K1 þ Kp :
svcl
; K2 ¼ K2 þ Kp :
svcl
;
1 þ Cxða 0Þ
svcl
K vp C3 1 þ Cxða 0Þ
svcl
K vp C3
:
C ða0 Þ
Kp xsvcl
Kp
Kpz ¼ 1
; K0p ¼ 1
: :


Cða0 ÞK C 1þ
Cða0 ÞK C
xsvcl vp 3 xsvcl vp 3
3.3 An Example Power System Installed with an SVC Stabilizer 111

: :

0
 Cxða 0Þ
Kvp C2 0
 Cxða 0Þ
Kvp C1
K3 ¼ K3 þ Kq :
svcl
; K4 ¼ K4 þ Kq :
svcl
;
1 þ Cxða 0Þ
svcl
K vp C3 1 þ Cxða 0Þ
svcl
K vp C3
:
C ða0 Þ
Kq xsvcl
Kq
Kqz ¼ 1
; K0q ¼ 1
: :


Cða0 ÞK C 1þ
Cða0 ÞK C
xsvcl vp 3 xsvcl vp 3
: :

 Cxða 0Þ
Kvp C1  Cxða 0Þ
Kvp C2
K05 ¼ K5 þ Kv svcl
: ; K06 ¼ K6 þ Kv svcl
: ;
1 þ Cxða 0Þ
svcl
Kvp C3 1 þ Cxða 0Þ
svcl
Kvp C3
:

Kv Kv Cxða 0Þ

Kvz ¼ 1
; K0v ¼ svcl
1
: :
Cða Þ
1 þ x 0 Kvp C3
Cða Þ
1 þ x 0 Kvp C3
svcl 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_ ¼ AX þ bDusvcs ð3:73Þ

where
3 2
2 0 3
Dd 6 K 70

6 7 6  p 7
6 Dx 7 6 M 7
6 7 6 K0q 7
6 DE0 7 6 7
X¼6 q 7; b ¼ 6  0 7;
6 7 6 T do 7
6 DE0fd 7 6 7
4 5 6 KA K0v 7
4 T 5
Dzsvc A

CU
2 3
0 xo 0 0 0
6 K 0
K 0
K 7
6  1
MD M 2
0  Mpz 7
6 M 7
6 0 0 7
6 K K Kqz 7
A ¼ 6  T0 4 0  03 1
T0do
 0 7
6 Tdo Tdo 7
6 7
do

6  KA K05 K K0
 TA 6  T1 K
 TA Kvz 7
4 TA 0
A A A
5
CA 0 CE 0 CZ
112 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

Linearization Eq. (3.63) is

ð1 þ sT2 Þð1 þ sT4 Þ


Dusvcs ¼ Ks ðDPt Þ ¼ Tsvcs ðsÞðDPt Þ
ð1 þ sT1 Þð1 þ sT3 Þ ð3:74Þ
¼ Tsvcs ðsÞðK01 Dd þ K02 DE0q þ Kpz Dz þ K0p Dusvcs Þ

Hence, the output equation of the system is as follows:

y ¼ cX þ dDusvcs ð3:75Þ

where
 
y ¼ DPt ; c ¼ K01 0 K02 0 Kpz ; d ¼ K0p

3.3.1.2 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 ¼ 1:0 p:u: This is achieved by an initial compensation from
the SVC at the operating condition, bsvc0 . This initial compensation can be calcu-
lated 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
2 sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3
 2ffi
Vs0 4 x
Psb0 5 ¼ 0:0377
sb
Qsb0 ¼ Vs0  V2b 
xsb Vs0

2sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3
 2
Vs0 4 x
Pts0  Vs0 5 ¼ 0:0377
ts
Qts0 ¼ V2t0 
xts Vs0

Hence, the initial compensation is calculated to be

Qsb0  Qts0
bsvc0 ¼ ¼ 0:0754
V2s0

This gives csvc0 ¼ 1  xsb bsvc0 ¼ 0:9774.


From Eq. (3.1), it can have

bsvc0 ¼ ½1  C(a0 Þ=Xsvcl  1=Xsvcc


3.3 An Example Power System Installed with an SVC Stabilizer 113

Thus, it can be obtained that at the steady-state operating condition,

2a0  sin 2a0


C(a0 Þ ¼  1 ¼ 0:0754
2p

Hence, a0 ¼ 42:0727 .
From the circuit model of the power system installed with the SVC in Fig. 3.5, it
can be calculated that

Is0 ¼ Vs0 jbSVC0 ¼ j0:0754; Its0 ¼ ðPt0  jQts0 Þ ¼ 0:5 þ j0:0377


Vs0
Isb0 ¼ Its0  Is0 ¼ 0:5 þ j0:0377; Vb0 ¼ Vs0  jIsb0 xsb ¼ 0:9887  j0:15
   
Psb0 þ jQb0 ¼ Isb0 Vb0 ¼ 0:4887  j0:1123

With the reference voltage, recalculation of the value of various variables is as


follows

Isb0 ¼ ðPt0  jQb0 Þ ¼ 0:5 þ j0:0377; V  s0 ¼ V b0 þ jXsbIsb0 ¼ 0:9887 þ j0:15


 b0
V
Is0 ¼ V
 s0 jbsvc0 ¼ 0:0113 þ j0:0746; Its0 ¼ Isb0 þ Is0 ¼ 0:4887 þ j0:1123
 t0 ¼ V
V  s0 þ jIts0 xts ¼ 0:955 þ j0:2966; E  Q0 ¼ V  t0 þ jxqIts0 ¼ 0:8652 þ j0:6876
d0 ¼ 38:4741
 
Id0 ¼ ðsin d0 þ j cos d0 ÞIts0 ¼ 0:2161; Eq0 ¼ EQ0 þ xd  xq Id0 ¼ 1:1483
 
E0q0 ¼ EQ0 þ x0d  xq Id0 ¼ 0:9646; Isbd0 ¼ ðsin d0 þ j cos d0 ÞIsb0 ¼ 0:2816
Isbq0 ¼ ðcos d0  j sin d0 ÞIsb0 ¼ 0:4149

From Eq. (3.7) it can have


xts
xlR ¼ xts þ ¼ 0:6069; x0dR ¼ x0d þ xlR ¼ 0:7569;
cSVC0
xdR ¼ xd þ xlR ¼ 1:6069; xqR ¼ xq þ xlR ¼ 1:4069

Thus, from Eqs. (3.9) and (3.12), parameters of the Heffron–Phillips model of
the example power system can be calculated to be

K1 ¼ 0:8765; K2 ¼ 0:8410; K3 ¼ 2:1229; K4 ¼ 0:7148; K5 ¼ 0:0473;


K6 ¼ 0:7475; Kp ¼ 0:0166; Kq ¼ 0:2990; Kv ¼ 0:0806
114 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

From Eqs. (3.71) and (3.72), it can have

CA ¼ 0:0358; CZ ¼ 0:4259; CE ¼ 0:7615; CU ¼ 0:1218;


K01 ¼ 0:8766; K02 ¼ 0:8394; Kpz ¼ 0:0175; K0p ¼ 0:0050; K03 ¼ 2:1514;
K04 ¼ 0:7134; Kqz ¼ 0:3158; K0q ¼ 0:0903; K05 ¼ 0:0476;
K06 ¼ 0:7398; Kvz ¼ 0:0852; K0v ¼ 0:0243:

State matrix of Eq. (3.73) thus is obtained to be


2 3
0 314:2 0 0 0
6 0:1461 0 0:1399 0 0:0029 7
6 7
A¼6
6 0:1415 0 0:4299 0:1983 0:0626 77
4 95:2539 0 1479:6 100:0 170:3097 5
0:0358 0 0:7614 0 0:4259

Eigenvalues of the state matrix are calculated to be

k1 ¼ 96:9595; k2;3 ¼ 0:1349 j6:7166


4 ¼ 3:2758; k
k 5 ¼ 0:3474

Hence, the electromechanical oscillation mode of the power system is


2;3 ¼ 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
8   9
< Kq þ Kv 1 +KAsT C2 Fsvc1 ðs)Kp þ K2 =
Fsvcs ðs) ¼ Fsvc2 ðs) Kp  A
: K3 þ sT0d0 þ K6 1 þKsT
A
þ C2 Fsvc1 ðs) Kq þ Kv 1 þKsT
A ;
A A

ð3:76Þ
Hence, from Fig. 3.17, the electric torque contribution from the SVC-based
stabilizer can be obtained to be
DTet ¼ Fsvcs ðs)Tsvcs ðs)(  DPt Þ ¼ Fsvcs ðs)Tsvcs ðs)(D þ sM)Dx ð3:77Þ
3.3 An Example Power System Installed with an SVC Stabilizer 115

K1

−ΔPt Δω ω0
1
-
D + sM s

Tsvc −s (s)
K2
Δu svc −s

C2 Fsvc1 (s) + Fsvc2 (s)

Δbsvc

Kp Kq Kv

1
- KA
K 3 + sTd0 ' 1 + sTA
-
+

K6

Fig. 3.18 Forward path of the SVC-based stabilizer

2;3 ¼ 0:1349 j6:7166. By taking


Oscillation modes of the power system are k
xs ¼ 6:7166, from Eqs. (3.67) and (3.76), it can have

svc1 ðjxs Þ ¼ 0:2899 þ j0:3040;


F svc2 ðjxs Þ ¼ 0:2696 þ j0:0171;
F

Fsvcs ðjxs Þ ¼ 0:0912  j0:0114

Take the transfer function of the SVC stabilizer to be

ð1 þ sT2 Þð1 þ sT4 Þ


Tsvcs ðs) ¼ Ks ; T1 ¼ 0:09 s:; T3 ¼ 0:09 s:
ð1 þ sT1 Þð1 þ sT3 Þ
116 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

The SVC stabilizer is to be designed to provide a damping torque


Dsvc Dx ¼ 1:6Dx, i.e., Dsvc ¼ 1:6. Parameters of the SVC-based stabilizer can be
set by using the phase compensation method to provide a pure damping torque

DTet ¼ Fsvcs ðjxs ÞðD þ jxs M)Tsvcs ðjxs ÞDx ¼ Dsvc Dx ð3:78Þ

The results of computation by the use of Eq. (3.78) are as follows

Ks ¼ K1 K2 ¼ 5:8540; T2 ¼ 1:7369 s; T4 ¼ 1:7369 s

Let

1 þ sT4 1 þ sT2
Dx1 ¼ Dx; Dusvcs ¼ K Dx1
1 þ sT3 1 þ sT1

Thus,
h
T4 K01
sDx1 ¼  T1 Dx1 þ T1  M Dd  DT M
4
 1 Dx
3
0
3
0

T4 K2 0 T4 Kpz T4 K
 M DEq  M Dzsvc  M p Dusvcs ð3:79Þ

sDusvc  s ¼ Ks1ð1þþsT
sT2 Þ Dx
1
1

From Eqs. (3.73) and (3.79), state equation of the closed-loop system can be
obtained to be
2 : 3 2 3
Dd Dd
6 : 7
6 Dx 7 6 Dx 7
6 0 7 6 7
6 DE_ q 7 6 DE0q 7
6 7 6 7
6
0
DE_ fd 7 ¼ Ac 6 DE0fd 7 ð3:80Þ
6 7 6 7
6 D_zsvc 7 6 Dzsvc 7
6 7 6 7
4 Dx_ 1 5 4 Dx1 5
Du_ svcs Dusvcs

where
2 3
0 314:1593 0 0 0 0 0
6 0:1461 0 0:1399 0 0:0029 0 0:0008 7
6 7
6 0:1415 0 0:4265 0:1983 0:0626 0 0:0179 7
6 7
Ac ¼ 6
6 95:2539 0 1479:6 100 170:3097 0 48:6868 7
7
6 0:0358 0 0:7615 0 0:4259 0 0:1218 7
6 7
4 0:7364 524:6004 0:2133 0:3170 0:0005 1:0411 0:0205 5
7:4877 5333:9 2:1683 3:2233 0:0056 4:7312 0:7913
3.3 An Example Power System Installed with an SVC Stabilizer 117

Fig. 3.19 Simulation of the δ /degree Without SVC stabilizer


example power system 42
without and with the SVC With SVC stabilizer
stabilizer installed 41

40

39

38

37

36

t/s
35
0 1 2 3 4 5 6 7 8 9 10

Eigenvalues of the state matrix of the closed-loop system are as follows

k1 ¼ 95:2487; 2;3 ¼ 0:6870 j6:6669;


k k4 ¼ 3:7484;
k5 ¼ 1:2281; k6 ¼ 0:2937; k7 ¼ 0:7921

The electromechanical oscillation mode of the power system is now k 2;3 ¼


0:6870 j6:6669 with sufficient 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 infinite-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 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

Fig. 3.20 Computational 45 Damping torque provided


results of the damping torque by the SVC stabilizer (p.u.)
provided by the SVC 40
stabilizer installed 35 Total damping torque

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 1

Table 3.1 Electromechanical oscillation mode


Load condition Pt0 (p.u.) 0.3 0.5 0.7
Electromechanical 0:1732 j5:4744 0:6870 j6:6669 1:0831 j6:0391
oscillation mode

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 confirm the analytical conclusion that
the effectiveness of the SVC stabilizer increases with the load condition of power
system.

Fig. 3.21 Simulation results δ /degree


Without the SVC
at Pt0 ¼ 0:3 27 stabilizer
With the SVC stabilizer
26

25

24

23

22

21

t/s
20
0 1 2 3 4 5 6 7 8 9 10
3.3 An Example Power System Installed with an SVC Stabilizer 119

Fig. 3.22 Simulation results δ /degree Without the SVC stabilizer


at Pt0 ¼ 0:7 58 With the SVC stabilizer

56

54

52

50

48

t/s
46
0 1 2 3 4 5 6 7 8 9 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 confirm 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 infinite busbar. From Fig. 3.23, it can be seen that
when the electric length of the line is longer (single transmission line connecting the

Damping torque provided by Single line


180 the SVC stabilizer(p.u.)

160

140

120

100
Double
80 lines

60

40

20
Pt0 (p.u.)
0
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Fig. 3.23 Computational results of the damping torque provided by the SVC stabilizer
120 3 Damping Torque Analysis of Thyristor-Based FACTS Stabilizers …

Damping torque provided Direct damping


12 torque
by the SVC stabilizer

10 Total damping
torque

2
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 infinite 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 ¼ 0:5; Vb0 ¼ 1:0; Vt0 ¼ 1:0; Vs0 ¼ 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):353–358
2. Smith OJM (1969) Power system transient control by capacitor switching. IEEE Trans Power
Apparatus Syst 88(1):28–35
3. Smith OJM, Webster RH (1971) Series capacitor switching to quench electromechanical
transients in power systems. IEEE Trans Power Apparatus Syst 90(2):427–433
4. Wang HF, Swift FJ (1996) Application of the Controllable Series Compensator in damping
power system oscillations. IEE Proc Part C 143(4):359–364
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):1–9
Chapter 4
Single-Machine Infinite-Bus Power
Systems Installed with VSC-Based
Stabilizers

4.1 Damping Torque Analysis of a Shunt VSC-Based


Stabilizer Installed in a Single-Machine Infinite-Bus
Power System

4.1.1 Extended Heffron–Phillips Model of a Single-Machine


Infinite-Bus Power System Installed with a Shunt
VSC-Based Stabilizer

4.1.1.1 A Shunt VSC Installed in a Single-Machine Infinite-Bus Power


System

Figure 4.1 shows the configuration 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 renew-
able power generation plant, such as a grid-connected photovoltaic (PV) or fuel cell
(FC) power plant. In case that Idc2 ¼ 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, ϕ (angle between Vc and Vs ), to charge or dis-
charge the capacitor. The amplitude of voltage at the AC terminal is regulated by

© Springer Science+Business Media New York 2016 121


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_4
122 4 Single-Machine Infinite-Bus Power Systems …

Fig. 4.1 Configuration of a transmission line Vs


shunt VSC-based unit

Is
xs

Vc
φ
m

Idc1
Idc

Cdc Idc2
Vdc

the modulation ratio m, i.e. Vc ¼ 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 configuration of a single-machine infinite-bus power
system, where a shunt VSC-based unit is installed at a location along the trans-
mission line. From Fig. 4.2, it can be obtained that

Vt ¼ jxts Its þ Vs
Vs ¼ jxs Is þ Vc ð4:1Þ
Vs  Vb ¼ jxsb ðIts  Is Þ

Fig. 4.2 A single-machine Vt Vs Vb


infinite-bus power system x ts x sb
installed with a VSC-based
unit
Its Isb

Is
xs

Vc
VSC-based unit
4.1 Damping Torque Analysis of a Shunt VSC-Based … 123

Hence

jxs Is þ Vc  Vb ¼ jxsb ðIts  Is Þ


ð4:2Þ
Vt ¼ jxts Its þ jxsb ðIts  Is Þ þ Vb

In the d–q coordinate of generator, as shown in Fig. 4.3, Eq. (4.2) gives

jxs ðisd þ jisq Þ þ Vc ðcos w þ j sin wÞ  Vb ðsin d þ j cos dÞ


¼ jxsb [(itsd  isd Þ þ j(itsq  isq Þ
ð4:3Þ
vtd þ jvtq
¼ jðxts þ xsb Þðitsd þ jitsq Þ  jxsb ðisd þ jisq Þ þ Vb ðsin d þ j cos dÞ

By equating the real and imaginary parts on both sides of Eq. (4.3), it can have

xs isq þ Vc cos w  Vb sin d ¼ xsb ðitsq  isq Þ


vtd ¼ ðxts þ xsb Þitsq þ xsb isq þ Vb sin d
ð4:4Þ
xs isd þ Vc sin w  Vb cos d ¼ xsb ðitsd  isd Þ
vtq ¼ ðxts þ xsb Þitsd  xsb isd þ Vb cos d

Because

vtd ¼ xq itsq
ð4:5Þ
vtq ¼ E0q  x0d itsd

Fig. 4.3 Phasor diagram of y q


system shown in Fig. 4.2

Vt

jxts Its
jxs Is
Vs
δ
φ
Vc jxsb Isb x
Vb
ψ

d
124 4 Single-Machine Infinite-Bus Power Systems …

From Eqs. (4.4) and (4.5), it can be obtained that


    
xsb xs  xsb itsq Vc cos w þ Vb sin d
¼
xq þ xts þ xsb xsb isq Vb sin d
   " # ð4:6Þ
xsb xs  xsb Itsd Vc sin w  Vb cos d
¼
x0d þ xts þ xsb xsb Isd E0q  Vb cos d

The above equation gives


   1  
itsq xsb xs  xsb Vc cos w þ Vb sin d
¼
isq xq þ xts þ xsb xsb Vb sin d
   1 " # ð4:7Þ
itsd xsb xs  xsb Vc sin w  Vb cos d
¼ 0
isd xd þ xts þ xsb xsb E0q  Vb cos d

The active power supplied by the generator can be expressed as

Pt ¼ vtd itsd þ vtq itsq ¼ xq itsq itsd þ ðE0q  x0q itsd Þitsq
ð4:8Þ
¼ E0q itsq þ ðxq  x0q Þitsd itsq

From Eqs. (2.32) and (4.5), it can have

Eq ¼ E0q þ ðxd  x0q Þitsd


qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð4:9Þ
Vt ¼ v2td þ v2tq ¼ ðxq itsq Þ2 þ ðE0q  x0q itsd Þ2

Magnitude of Vc is mkVdc . In the d–q coordinate of generator as shown in Fig. 4.3,


its phase is w. Hence,

Vc ¼ mkVdc ðcos w þ j sin wÞ ¼ mkVdc \w ð4:10Þ

Active power received by the DC capacitor of VSC-based unit is

Vdc Idc1 ¼ isd vcd þ isq vcq ¼ isd mkVdc cos w þ isq mkVdc sin w ð4:11Þ

The above equation gives


Idc ¼ isd mk cos w þ isq mk sin w ð4:12Þ

Dynamic equation on the DC side of the converter is

1 1 Idc2
V_ dc ¼ ðIdc1 þ Idc2 Þ ¼ ðisd mk cos w þ isq mk sin wÞ þ ð4:13Þ
Cdc Cdc Cdc
4.1 Damping Torque Analysis of a Shunt VSC-Based … 125

From the phasor diagram of Fig. 4.3, it can have


vsq
w ¼ tan1 / ð4:14Þ
vsq

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 infinite-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 Heffron–Phillips Model

Linearization of Eq. (4.7) is


   
Ditsq c11 c12
¼
Disq c21 c22
 
m0 k cos w0 DVdc  kVdc0 cos w0 Dm þ m0 kVdc0 sin w0 Dw þ Vb cos d0 Dd

Vb cos d0 Dd
   
Ditsd d11 d12
¼
Disd d21 d22
" #
m0 k sin w0 DVdc þ kVdc0 sin w0 Dm þ m0 kVdc0 cos w0 Dw þ Vb sin d0 Dd

DE0q þ Vb sin d0 Dd
ð4:15Þ
The second equation in Eq. (4.1) is
vsd þ jvsq ¼ jxs ðisd þ jisq Þ þ mkVdc cos w þ jmkVdc sin w

Hence, it can have


vsd ¼ xs isq þ mkVdc cos w
vsq ¼ xs isd þ mkVdc sin w

By using Eq. (4.15), linearization of above equations can be obtained as

Dvsd ¼ ðxs þ 1Þm0 k cos w0 DVdc þ ðxs þ 1ÞkVdc0 cos w0 Dm


 ðxs þ 1Þm0 kVdc0 sin w0 Dw  xs Vb cos d0 Dd
¼ ad1 DVdc þ ad2 Dm þ ad3 Dw þ ad4 Dd
Dvsq ¼ ðxs þ 1Þm0 k sin w0 DVdc þ ðxs þ 1ÞkVdc0 sin w0 Dm
þ ðxs þ 1Þm0 kVdc0 cos w0 Dw þ Vb sin d0 D
¼ aq1 DVdc þ aq2 Dm þ aq3 Dw þ aq4 Dd
126 4 Single-Machine Infinite-Bus Power Systems …

By using above equations, linearization of Eq. (4.14) can be obtained as

Dw ¼ a01 DVdc þ a02 Dm þ a03 Dw þ a04 Dd  D/

where

1
a01 ¼ ðvsd0 a1q  vsq0 a1d Þ
v2sd0 þ v2sq0
1
a02 ¼ 2 ðvsd0 a2q  vsq0 a2d Þ
vsd0 þ v2sq0
1
a03 ¼ 2 ðvsd0 a3q  vsq0 a3d Þ
vsd0 þ v2sq0
1
a04 ¼ 2 ðvsd0 a4q  vsq0 a4d Þ
vsd0 þ v2sq0

That is

a01 a02 a04 1


Dw ¼ 0 DVdc þ 0 Dm þ Dd  D/
1  a3 1  a3 1  a03 1  a03
¼ a1 DVdc þ a2 Dm þ a3 Dd þ a4 D/ ð4:16Þ

By using Eqs. (4.15) and (4.16), linearization of Eqs. (4.8) and (4.9) can be
obtained as

DPt ¼ K01 Dd þ K2 DE0q þ K0pdc DVdc þ K0pm Dm þ K0pw Dw


¼ ðK01  a3 K0pw ÞDd þ K2 DE0q þ ðK0pdc  a1 K0pw ÞDVdc
þ ðK0pm  a2 K0pw ÞDm þ a4 K0pw D/
¼ K1 Dd þ K2 DE0q þ Kpdc DVdc þ Kpm Dm þ Kpw D/
DEq ¼ K04 Dd þ K3 DE0q þ K0qdc DVdc þ K0qm Dm þ K0qw Dw
¼ ðK04  a3 K0qw ÞDd þ K3 DE0q þ ðK0qdc  a1 K0qw ÞDVdc
ð4:17Þ
þ ðK0qm  a2 K0qw ÞDm þ a4 K0qw D/
¼ K4 Dd þ K3 DE0q þ Kqdc DVdc þ Kqm Dm þ Kqw D/
DVt ¼ K05 Dd þ K6 DE0q þ K0vdc DVdc þ K0vm Dm þ K0vw Dw
¼ ðK5  a3 K0vw ÞDd þ K6 DE 0 q þ ðK0vdc  a1 K0vw ÞDVdc
þ ðK0vm  a2 K0vw ÞDm þ a4 K0vw D/
¼ K5 Dd þ K6 DE0q þ Kvdc DVdc þ Kvm Dm þ Kvw D/
4.1 Damping Torque Analysis of a Shunt VSC-Based … 127

where

C0 ¼ m0 k
h  i
K01 ¼ E0q0  itsd0  x0d  xq ðc11 þ c12 ÞVb cos d0
 
 itsq0  x0d  xq ðd11 þ d12 ÞVb sin d0
 
K2 ¼ itsq0  itsq0  x0d  xq d12
h  i  
K0pdc ¼ c11 C0 cos w0 E0q0  itsd0  x0d  xq  d11 C0 sin w0 Itsq0  x0d  xq
h  i  
K0pm ¼ c11 kVdc0 cos w0 E0q0  itsd0  x0d  xq  d11 kVdc0 sin w0 Itsq0  x0d  xq
h  i  
K0pw ¼ c11 C0 Vdc0 sin w0 E0q0  itsd0  x0d  xq  d11 C0 Vdc0 cos w0 Itsq0  x0d  xq
K4 ¼ ðd11 þ d12 ÞVb sin d0 ðxd  x0d Þ
K03 ¼ d12 ðxd  x0d Þ þ 1
K0qdc ¼ d11 C0 sin w0 ðxd  x0d Þ
K0qm ¼ d11 kVdc0 sin w0 ðxd  x0d Þ
K0qw ¼ d11 C0 Vdc0 cos w0 ðxd  x0d Þ
 
0 vtd0 Xq ðc11 þ c12 ÞVb cos d0  vtq0 x0d ðd11 þ d12 ÞVb sin d0
K5 ¼
Vt0
 
vtq0 1  x0d d12
K6 ¼
Vt0
 
0 v td0 xq c11 C0 cos w0  vtq0 x0d d11 C0 sin w0
Kvdc ¼
Vt0
 
0 v td0 xq kc 11 V dc0 cos w0  vtq0 x0d d11 kVdc0 sin w0
Kvm ¼
Vt0
 
0 vtd0 xq c11 C0 Vdc0 sin w0  vtq0 x0d d11 C0 Vdc0 cos w0
Kvw ¼
Vt0

Linearization of Eq. (4.13) is

Cdc DV_ dc ¼ ðisd0 k cos w0 þ isq0 k sin w0 ÞDm


þ m0 kðisd0 sin w0 þ isq0 cos w0 ÞDw
þ m0 kðcos w0 Disd þ sin w0 Disq Þ þ DIdc2
128 4 Single-Machine Infinite-Bus Power Systems …

By using Eqs. (4.15) and (4.16), it can have

1
DV_ dc ¼ K07 Dd þ K8 DE0q þ K09 DVdc þ K0dm Dm + K0dw Dw þ DIdc2
Cdc
¼ ðK07  a3 K0dw ÞDd þ K8 DE0q þ ðK09  a1 K0dw ÞDVdc
1
þ ðKdm  a2 K0dw ÞDm + a4 K0dw Df + DIdc2
Cdc
1
¼ K7 Dd þ K8 DE0q þ K9 DVdc þ Kdm Dm + Kdw D/ þ DIdc2 ð4:18Þ
Cdc

where

½m0 k cos w0 ðd21 þ d22 ÞVb sin d0 þ m0 k sin w0 ðc21 þ c22 ÞVb cos d0 
K07 ¼
Cdc
m0 k cos w0 d22
K8 ¼
C
 2 2 dc 
0 m k cos w0 d21 sin w0  m20 k2 sin w0 c21 cos w0
K9 ¼ 0
Cdc
1
K0dm ¼ ðkisd0 cos w0 þ kisq0 sin w0 þ m0 k2 cos w0 d21 Vdc0 sin w0
Cdc
 m0 k2 sin w0 c21 Vdc0 cos w0 Þ
1
K0dw ¼ ½m0 kðisd0 sin w0  isq0 cos w0 Þ þ m20 k2 cos w0 d21 Vdc0 cos w0
Cdc
þ m20 k2 sin w0 c21 Vdc0 sin w0 

By substituting Eq. (4.17) into the linearized differential equation of syn-


chronous generator of Eq. (2.37) (for the simplicity of expression, the PSS is not
considered), the extended Heffron–Phillips model of the single-machine infinite-bus
power system installed with the shunt VSC-based unit can be obtained as

Dd_ ¼ xo Dx
1
Dx_ ¼ ðK1 Dd  DDx  K2 DE0 q  Kpdc DVdc  Kpm Dm  Kpw D/Þ
M
1
DE_ 0q ¼ 0 ðK4 Dd  K3 DE0 q þ DE0fd þ Kqdc DVdc þ Kqm Dm þ Kqw D/Þ
Td0
1 KA
DE_ 0fd ¼  DE0fd  ðK5 Dd þ K6 DE0 q þ Kvdc DVdc þ Kvm Dm þ Kvw D/Þ
TA TA
1
DV_ dc ¼ K7 Dd þ K8 DE0 q þ K9 DVdc þ Kdm Dm þ Kdw D/ þ DIdc2
Cdc
ð4:19Þ
4.1 Damping Torque Analysis of a Shunt VSC-Based … 129

The model is shown in Figs. 4.4 and 4.5.

ΔIdc2

1
Cdc
ΔE q' K8
1
+ ΔVdc
Δδ K7 s − K9

K dm K dψ

Δm Δφ

Fig. 4.4 Extended Heffron–Phillips model—part of the VSC-based unit

K 1 − K Pψ

1 Δω ω0 Δδ
- sM + D s

K 4 − K qψ K 5 − K vψ
K2 ΔVdc
[K pdc K pψ K pm ] [Δψ ] [K vdc K vψ K vm ]
Δm

[K qdc K qψ K qm ]

1 KA
ΔE q ' Td0 's + K 3 - + 1 + sTA -
K6

Fig. 4.5 Extended Heffron–Phillips model—part of the power system


130 4 Single-Machine Infinite-Bus Power Systems …

4.1.2 Damping Torque Analysis of Shunt VSC-Based


Stabilizer Installed in Single-Machine Infinite-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 reg-
ulating 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 ðsÞ, respectively, and only the
stabilizer’s function is considered, it can have Dm ¼ Tvsc ðsÞDy. At the angular
oscillation frequency of the power system, xs , stabilizing signal of the VSC-based
reactive power stabilizer Dm ¼ Tvsc ðjxs ÞDy can always be decomposed into a
synchronizing and damping torque, that is

Dm ¼ Tvsc ðjxs ÞDy ¼ Kvscd Dx þ Kvscs Dd ð4:20Þ

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 ¼ Krs Dx ð4:21Þ

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

K9
DTrsd ¼ Kpm  Kpdc Kdm Krs Dx ð4:22Þ
x2s þ K29

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
4.1 Damping Torque Analysis of a Shunt VSC-Based … 131

D/ ¼ Kas Dx ð4:23Þ

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

K9
DTasd ¼ Kpw  Kpdc Kdw 2 Kas Dx ð4:24Þ
xs þ K29

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

From Fig. 4.2, it can have

jxsb xsb Vb
VS ¼ x Its þ x Vc þ ð4:25Þ
1 þ xs sb
xs ð1 þ xs Þ
sb
1 þ xxsbs

That gives

Vt ¼ jxts Its þ Vs

xs xsb xsb xs ð4:26Þ


¼ j xts þ Its þ Vc þ Vb ¼ jxIts þ Va
xs þ xsb xs þ xsb xs þ xsb

where

xs xsb
x¼ xts þ
xs þ xsb
xsb xs
Va ¼ Vc þ Vb ¼ aVc þ bVb
xs þ xsb xs þ xsb

Hence, from Eq. (4.26), the active power supplied by the generator can be
expressed as

E0q Va 0 V2a ðxq  x0d Þ


Pt ¼ sin d  sin 2d0 ð4:27Þ
x0dR 2 x0dR xqR

where d0 is the angle between E0q and Va and x0dR ¼ x þ x0d ; xqR ¼ x þ xq . From the
phasor diagram of Fig. 4.6, it can have
132 4 Single-Machine Infinite-Bus Power Systems …

Fig. 4.6 Phasor diagram for q


Eq. (4.27)

E’
q

δ Va
δ aVc
bVb ψ
d
d

Va sin d0 ¼ bVb sin d þ aVc cos w


ð4:28Þ
Va cos d0 ¼ bVb cos d þ aVc sin w

Thus, from Eqs. (4.27) and (4.28), the active power supplied by the generator can
be expressed alternatively to be

E0q
Pt ¼ ðbVb sin d þ aVc cos wÞ
x0dR
ðxq  x0d Þ
 0 ðbVb sin d þ aVc cos wÞðbVb cos d þ aVc sin wÞ ð4:29Þ
xdR xqR
E0q
 0 ðbVb sin d þ aVc cos wÞ
xdR

where Vc ¼ mkVdc . Equation (4.29) is the explicit expression of the active power
supplied by the generator in the single-machine infinite-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
Kpm ¼ a4 ¼ a4 akVdc0 cos w0
@m 0

@Pt
Kpw ¼ a4 ¼ a4 akm0 Vdc0 sin w0 ð4:30Þ
@w 0

@Pt
Kpdc ¼ a4 ¼ a4 akm0 cos w0
@Vdc 0
4.1 Damping Torque Analysis of a Shunt VSC-Based … 133

From Eqs. (4.22) and (4.30), the damping torque contributed by the VSC-based
reactive power stabilizer is obtained as

K9
DTrsd ¼ a4 ak cos w0 Vdc0  m0 Kdm Krs Dx ð4:31Þ
x2s þ K29

From Eqs. (4.24) and (4.30), the damping torque contributed by the VSC-based
active power stabilizer is derived as

K9
DTasd ¼ a4 akm0 Vdc0 sin w0 þ cos w0 Kdw Kas Dx ð4:32Þ
x2s þ K29

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 infinite-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 Infinite-Bus Power System

4.2.1 Damping Torque Analysis of a SSSC Stabilizer


Installed in a Single-Machine Infinite-Bus Power
System

4.2.1.1 A Single-Machine Infinite-Bus Power System Installed


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 capac-
itive 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 4 Single-Machine Infinite-Bus Power Systems …

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 infinite-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 com-
pensation level can be controlled dynamically by regulating the magnitude of Vinv .
Mathematical description of the SSSC is

Vt It xt Vb

Pt
+ V sssc -
I ts
V1 + V inv - V2
x sct
x ts x sb
Pts

m
VSC
φ

Idc Vdc

+ -

Fig. 4.7 A single-machine infinite-bus power system with a SSSC

V inv V inv

V1 V2
jx cst I ts jx cst I ts
V2 V1

I ts I ts
SSSC in inductive mode SSSC in capacitive mode

Fig. 4.8 Phasor diagram of operation of the SSSC


4.2 Damping Function of a Stabilizer Added on a Static … 135

Its ¼ itsd þ jitsq ¼ Its \u


Vinv ¼ mkVdc ðcos / þ j sin /Þ ¼ mkVdc \/; u þ / ¼ 90 ð4:33Þ
dVdc Idc mk
¼ ¼ ðitsd cos / þ itsq sin /Þ
dt Cdc Cdc

where k is the fixed ratio between the inverter’s 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 ¼ jxdc Its ð4:34Þ

From the circuit equation of power system of Fig. 4.7, it can have

Vt  Vb ¼ j(xts þ xsb þ xsct ÞIts þ Vinv ¼ jxt It ð4:35Þ

In the d–q coordinate, the above equation becomes

ðxts þ xsb þ xsct Þitsq þ j(xts þ xsb þ xsct Þitsd  xdc itsq þ jxdc itsd
ð4:36Þ
¼ xt itq þ jxt itd

Hence,
xts þ xsb þ xsct þ xdc
itq ¼ itsq
xt
ð4:37Þ
xts þ xsb þ xsct þ xdc
itd ¼ itsd
xt

Considering the voltage equation of generator of Eq. (4.35), it can have

Vt  j(xts þ xsb þ xsct ÞIts ¼ xq ðitsq þ itq Þ þ j½E0q  x0d ðitsd þ itd Þ  jðxts þ xsb þ xsct ÞIts



xts þ xsb þ xsct þ xdc xts þ xsb þ xsct þ xdc
¼ xq 1 þ itsq þ j E0q  x0d 1 þ itsd
xt xt
þ ðxts þ xsb þ xsct Þitsq  j(xts þ xsb þ xsct Þitsd
¼ Vinv þ Vb ¼ xdc itsq þ jxdc itsd þ Vb sin d þ jVb cos d ð4:38Þ

From the above equation, it can be obtained that

Vb sin d
itsq ¼
xts þ xsb þ xsct þ cs xq þ xdc
ð4:39Þ
E0q  Vb cos d
itsd ¼
xts þ xsb þ xsct þ cs x0d þ xdc
136 4 Single-Machine Infinite-Bus Power Systems …

where
xts þ xsb þ xsct þ xdc
cs ¼ 1 þ
xt

Thus,

Pt ¼ E0q ðitq þ itsq Þ þ ðxq  x0d Þðitd þ itsd Þðitq þ itsq Þ


¼ E0q cs itsq þ ðxq  x0d Þc2s itsd itsq
Eq ¼ E0q þ ðxd  x0d Þðitd þ itsd Þ ¼ E0q þ ðxd  x0d Þcs itsd
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð4:40Þ
Vt ¼ ½E0q  x0d ðitd þ itsd Þ2 þ ½xq ðitq þ itsq Þ2
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
¼ ðE0q  x0d cs itsd Þ2 þ ðxq cs itsq Þ2

From Eqs. (4.33) and (4.34), it can have

Vinv ¼ jxdc Its ¼ xdc itsq þ jxdc itsd ¼ mkVdc ðcos / þ j sin /Þ ð4:41Þ

The above equation gives


xdc itsq
cos / ¼
mkVdc
ð4:42Þ
xdc itsd
sin / ¼
mkVdc

By using Eq. (4.42), dynamic equation in Eq. (4.33) is converted to the following
equivalent form
dVdc
¼0 ð4:43Þ
dt

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 Heffron–Phillips Model of the Single-Machine


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

Linearization of Eq. (4.40) is

DPt ¼ K1 Dd þ K2 DE0q þ Kpx Dxdc


DEq ¼ K4 Dd þ K3 DE0q þ Kqx Dxdc ð4:44Þ
DVt ¼ K5 Dd þ K6 DE0q þ Kvx Dxdc
4.2 Damping Function of a Stabilizer Added on a Static … 137

From Eq. (4.41), it can have

xdc Its ¼ mkVdc ð4:45Þ

Hence,

Its0 Dxdc þ xdc0 DIts ¼ kVdc0 Dm ð4:46Þ

Because only the damping control function will be considered, it can be assumed
that xdc0 ¼ 0 as far as the damping control function at the steady-state operation is
concerned. Hence, Eq. (4.44) becomes

DPt ¼ K1 Dd þ K2 DE0q þ Kpm Dm


DEq ¼ K4 Dd þ K3 DE0q þ Kqm Dm ð4:47Þ
DVt ¼ K5 Dd þ K6 DE0q þ Kvm Dm

where

Kpx kVdc0
Kpm ¼
Its0
Kqx kVdc0
Kqm ¼
Its0
Kvx kVdc0
Kvm ¼
Its0

By substituting Eq. (4.47) into Eq. (2.39) (without the PSS being considered), the
extended Heffron–Phillips model of the single-machine infinite-bus power system
with the SSSC installed can be obtained as

Dd_ ¼ xo Dx
1
Dx_ ¼ ½K1 Dd  DDx  K2 DE0q  Kpm Dm
M
1 ð4:48Þ
DE_ 0q ¼ 0 ðK4 Dd  K3 DE0q þ DE0fd  Kqm DmÞ
Td0
1 KA
DE_ 0fd ¼  DE0fd  ðK5 Dd þ K6 DE0q þ Kvm DmÞ
TA TA

The model is shown in Fig. 4.9.


Active power supplied by the generator can be expressed as

Pt ¼ E0q ðitsq þ itq Þ þ ðxq  x0d Þðitsd þ itd Þðitsq þ itq Þ


ð4:49Þ
¼ E0q cs itsq þ ðxq  x0d Þc2s itsd itsq
138 4 Single-Machine Infinite-Bus Power Systems …

K1

1 ω0
- D + sM s

Δm
K2 K4 K5
K pm K qm K vm

1
- - KA
K 3 + sTd0 ' + 1 + sTA
-

K6

Fig. 4.9 Extended Heffron–Phillips model of single-machine infinite-bus power system installed
with a SSSC stabilizer

By using Eq. (4.39), it can have

cs E0q Vb c2s V2b ðxq  x0d Þ


Pt ¼ sin d  sin 2d ð4:50Þ
x0dR 2 x0dR xqR

where

x0dR ¼ cs x0d þ xts þ xsb þ xsct þ xdc


xqR ¼ cs xq þ xts þ xsb þ xsct þ xdc

From Eq. (4.50), it can be defined that

cs E0q0 Vb c2 V2 ðxq  x0d Þ


Pt0 ¼ 0 sin d0  s b 0 sin 2d0 ¼ Pt10  Pt20
xdR 2 xdR xqR
cs E0q0 Vb
Pt10 ¼ sin d0 ð4:51Þ
x0dR
c2s V2b ðxq  x0d Þ
Pt20 ¼ sin 2d0
2 x0dR xqR
4.2 Damping Function of a Stabilizer Added on a Static … 139

According to Eqs. (4.41) and (4.51), from Eq. (4.50), it can have
0 0
xd xq xd
@Pt 1 xt þ 1 2 xt þ 1 xt þ 1
Kpx ¼ ¼ Pt10  Pt10  Pt20 þ Pt20 þ Pt20
@xdc 0 xt cs x0dR xt c s xqR x0
0 dR
xd xq
1 1 xL þ 1 xL þ 1
¼ ðPt10  Pt20 Þ  Pt20  ðPt10  Pt20 Þ þ Pt20
xt cs xt c s x0dR xqR

1 x0 þ xt 1 xq þ xt
¼  d 0 Pt0   Pt20
xt c s xt xdR xt c s xt xqR
x0  cs x0d  cs xt xqR  cs xq  xt cs
¼ dR Pt0  Pt20
xt cs x0dR xt cs xqR
1 1
¼ Pt0 þ Pt20
cs x0dR cs xqR
ð4:52Þ

From Eqs. (4.47) and (4.52), it can be obtained that


kVdc0 1 1
Kpm ¼ Pt0  Pt20 ð4:53Þ
cs Its0 x0dR xqR

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 infinite-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 4 Single-Machine Infinite-Bus Power Systems …

Δm
K2

K pm K qm K vm

1
- KA
K 3 + sTd0 ' + 1 + sTA
-

K6

Fig. 4.10 Forward path of stabilizing signal of SSSC stabilizer installed in a single-machine
infinite-bus power system
1
K3 þ sT0d0
Fsssc ðsÞ ¼ Kpm  K2 Kqm
1 þ K61 KA
K3 þ sT0d0 1 þ sTA
1 KA
K3 þ sT0d0 1 þ sTA ð4:54Þ
 K2 Kvm
1 þ K6 1 KA
K3 þ sT0d0 1 þ sTA
K2 ½Kqm ð1 þ sTA Þ þ Kvm KA 
¼ Kpm 
ðK3 þ sT0d0 Þð1 þ sTA Þ þ K6 KA

Consider the case that the feedback signal and transfer function of the SSSC
stabilizer are Dx and Tsssc ðs), respectively, that is
Dm ¼ Tsssc ðsÞDx ð4:55Þ
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 ðjxs ÞTsssc ðjxs Þ ¼ Dsssc ð4:56Þ
In a more general case, the state-space representation of linearized model of the
single-machine infinite-bus power system with the SSSC stabilizer installed can
always be arranged in the following form:
2 3 2 3 2
32 3
Dd_ 0 x0 0 Dd 0
6 7 6 76 7 6 7
4 Dx_ 5 ¼ 4 k d A23 54 Dx 5 þ 4 B2 5Dm
x_ A31 A32 A33 x B3
2 3 ð4:57Þ
Dd
6 7
Dy ¼ ½ C1 C2 C3 4 Dx 5 þ dm Dm
x
4.2 Damping Function of a Stabilizer Added on a Static … 141

Equation (4.57) can be written as

Dd_ ¼ x0 Dx
Dx_ ¼ kDd  dDx  A23 x  B2 Dm
ð4:58Þ
x_ ¼ A31 Dd þ A32 Dx þ A33 x þ B3 Dm
Dy ¼ C1 Dd þ C2 Dx þ CT3 x þ dm Dm

From Eq. (4.58), it can have

sDd ¼ x0 Dx
sDx ¼ kDd  dDx  A23 x  B2 Dm
ð4:59Þ
sx ¼ A31 Dd þ A32 Dx þ A33 x þ B3 Dm
Dy ¼ C1 Dd þ C2 Dx þ CT3 x þ dm Dm

Hence,

s
x ¼ ðsI  A33 Þ1 A31 þ A32 Dd þ ðsI  A33 Þ1 B3 Dm ð4:60Þ
x0

From Eqs. (4.59) and (4.60), it can be obtained that

sDd ¼ x0 Dx


s s
sDx ¼  k þ d þ A23 ðsI  A33 Þ1 A31 þ A32 Dd
x0 x0
 ½A23 ðsI  A33 Þ1 B3 þ B2 Dm ¼ KðsÞDd  Kc ðsÞDm


s
Dy ¼ C1 Dd þ C2 Dx þ CT3 ðsI  A33 Þ1 A31 þ A32 Dd þ ðsI  A33 Þ1 B3 Dm
x0
h x i
1 x0
A31 þ A32 Dx þ ½dm þ CT3 ðsI  A33 Þ1 B3 Dm
0
¼ C1 þ C2 þ C3 ðsI  A33 Þ
T
s s
¼ Ko ðsÞDx þ Kil ðsÞDm
ð4:61Þ

where

s s
KðsÞ ¼ k þ d þ AT23 ðsI  A33 Þ1 A31 þ A32
x0 x0
x
0 1 x0
Ko ðsÞ ¼ C1 þ C2 þ C3 ðsI  A33 Þ
T
A31 þ A32
s s
Kc ðsÞ ¼ AT23 ðsI  A33 Þ1 B3 þ B2
Kil ðsÞ ¼ CT3 ðsI  A33 Þ1 B3 þ dm

Hence, according to Eq. (4.61), the system is shown in Fig. 4.11.


142 4 Single-Machine Infinite-Bus Power Systems …

Fig. 4.11 Closed-loop Δδ ω0


system installed with the K(s)
s
SSSC stabilizer
Electromechanical oscillation loop
-
1 Δω
s
-

K c (s) K o (s)

+
+
K il (s)
y
Δm

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

Kc ðjxs ÞKo ðjxs ÞTsssc ðjxs Þ


DTet ¼ Dx ð4:62Þ
1  Kil ðjxs ÞTsssc ðjxs Þ

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 ¼ Dsssc Dx, that is

Kc ðjxs ÞKo ðjxs ÞTsssc ðjxs Þ


¼ Dsssc ð4:63Þ
1  Kil ðjxs ÞTsssc ðjxs Þ

Hence,

Dsssc ¼ ½Kc ðjxs ÞKo ðjxs Þ þ Dsssc Kil ðjxs ÞTsssc ðjxs Þ
ð4:64Þ
¼ Fsssc ðjxs ÞTsssc ðjxs Þ

According to Eq. (4.64), the phase compensation method can be used to design the
SSSC stabilizer by setting

Dsssc
h ¼ u; Tsssc ¼ ð4:65Þ
Fsssc

where Tsssc ðjxs Þ ¼ Tsssc \hFsssc ðjxs Þ ¼ Fsssc \u.


4.3 Damping Function of a Unified Power Flow Controller … 143

4.3 Damping Function of a Unified Power Flow Controller


(UPFC) Installed in a Single-Machine Infinite-Bus
Power System

4.3.1 Mathematical Model of a Single-Machine Infinite-Bus


Power System Installed with a UPFC

4.3.1.1 Dynamic Model of a UPFC

Figure 4.12 shows the configuration of a single-machine infinite-bus power system


installed with a unified 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 ¼ me ke Vdc \wes


ð4:66Þ
Vbs ¼ mb kb Vdc \wbs

Vt Vb
Itb x tb

Pt Vbt
I te +
Vet I bs Vbs
+
x te x bt
x bs BT

Pte Ies VSC − E VSC − B


ET
x es
Iedc I bdc
+
Ves Cdc

m e δe UPFC m b δb

Fig. 4.12 A single-machine infinite-bus power system installed with a UPFC


144 4 Single-Machine Infinite-Bus Power Systems …

y q y q

Vet Vbt

δ jx es Ies jx bs I bs
δ
δe δb
Ves Vbs
x x

ψ es ψ bs
d d

Fig. 4.13 Phase diagram of shunt and series SVS

where wes and wbs are the angle between d-axis of d–q coordinate of the synchronous
generator and voltage Ves and Vbs , respectively, as shown in Fig. 4.13, which gives

Vet ¼ jxes Ies þ Ves


ð4:67Þ
Vbt ¼ jxbs Ibs þ Vbs

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 ¼ ReðVes Ies Þ and Pbs ¼ ReðVbs Ibs Þ, where ReðÞ 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
Pes ¼ sin de ¼ sin de
xes xe
ð4:68Þ
Vbs Vbt mb kb Vdc Vbt
Pbs ¼ sin db ¼ sin db
xbs xbs

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
Pes ¼ sin de ¼ Iedc Vdc
xes
ð4:69Þ
mb kb Vdc Vbt
Pbs ¼ sin db ¼ Ibdc Vdc
xbs
4.3 Damping Function of a Unified Power Flow Controller … 145

From Eq. (4.69), it can have

me ke Vet
Iedc ¼ sin de
xes
ð4:70Þ
mb kb Vbt
Ibdc ¼ sin db
xbs

Alternatively, in the d–q coordinate as shown in Fig. 4.13,

Pes ¼ vesd iesd þ vesq iesq ¼ iesd me ke Vdc cos wes þ iesq me ke Vdc sin wes ¼ Iedc Vdc
Pbs ¼ vbsd ibsd þ vbsq ibsq ¼ ibsd mb kb Vdc cos wbs þ ibsq mb kb Vdc sin wbs ¼ Ibdc Vdc
ð4:71Þ

From Eq. (4.71), it can have

Iedc ¼ iesd me ke cos wes þ iesq me ke sin wes


ð4:72Þ
Ibdc ¼ ibsd mb kb cos wbs þ ibsq mb kb sin wbs

Hence, from Eq. (4.70), dynamic equation of the UPFC can be obtained as

Iedc þ Ibdc 1 me ke Vet mb kb Vbt


V_ dc ¼ ¼ sin de þ sin db ð4:73Þ
Cdc Cdc xes xbs

Alternatively, from Eq. (4.72), it can have

1
V_ dc ¼ ðiesd me ke cos wes þ iesq me ke sin wes þ ibsd mb kb cos wbs þ ibsq mb kb sin wbs Þ
Cdc
ð4:74Þ

where
vetq
wes ¼ tan1  de
vetd
vbtq ð4:75Þ
wbs ¼ tan1  db
vbtd
146 4 Single-Machine Infinite-Bus Power Systems …

4.3.1.2 Nonlinear Model of the Single-Machine Infinite-Bus Power


System Installed with the UPFC

From Fig. 4.12, it can have

Vt ¼ jxtb Itb þ Vb
Vt ¼ jxte Ite þ Vet ð4:76Þ
Vet ¼ Vbt þ jxbt Ibs þ Vb

which can be expressed in the d–q coordinate as

vtd þ jvtq ¼ jxtb ðitbd þ jitbq Þ þ Vb sin d þ jVb cos d


¼ xtb itbq þ Vb sin d þ jxtb itbd þ jVb cos d
ð4:77Þ
vtd þ jvtq ¼ jxte ðiesd þ ibsd þ jiesq þ jibsq Þ þ vetd þ jvetq
vetd þ jvetq ¼ vbtd þ jvbtq þ jxbt ibsd  xbt ibsq þ Vb sin d þ jVb cos d

From the first equation in Eq. (4.77), it can be obtained that

vtq  Vb cos d
itbd ¼
xtb
ð4:78Þ
Vb sin d  vtd
itbq ¼
xtb

Because

vtd ¼ xq ðiteq þ itbq Þ


ð4:79Þ
vtq ¼ E0q  x0d ðited þ itbd Þ

From Eqs. (4.78) and (4.79), it can have

xq xtb iteq þ xq Vb sin d


vtd ¼
xq þ xtb
0 ð4:80Þ
xtb Eq  x0d xtb ited þ x0d Vb cos d
vtq ¼
x0d þ xtb
4.3 Damping Function of a Unified Power Flow Controller … 147

The second and third equations in Eq. (4.77) give

vtd þ jvtq ¼ jxte ðiesd þ ibsd þ jiesq þ jibsq Þ  xes iesq þ vesd þ jxes iesd þ jvesq
 xes iesq þ vesd þ jxes iesd þ jvesq
¼ xbs ibsq þ vbsd þ jxbs ibsd þ jvbsq þ jxbt ibsd  xbt ibsq þ Vb sin d þ jVb cos d

That is

vtd ¼ xte ðiesq þ ibsq Þ  xes iesq þ vesd


vtq ¼ xte ðiesd þ ibsd Þ þ xes iesd þ vesq
ð4:81Þ
xes iesq þ vesd ¼ xbs ibsq þ vbsd  xbt ibsq þ Vb sin d
xes iesd þ vesq ¼ xbs ibsd þ vbsq þ xbt ibsd þ Vb cos d

From Eqs. (4.80) and (4.81), it can be obtained that

½xq xtb þ xte ðxq þ xtb Þ þ xes ðxq þ xtb Þiesq


þ ½xq xtb þ xte ðxq þ xtb Þibsq ¼ ðxq þ xtb Þvesd  xq Vb sin d
½xte ðx0d þ xtb Þ þ xes ðx0d þ xtb Þ þ x0d xte iesd
þ ½xte ðx0d þ xtb Þ þ x0d xtb ibsd ¼ xtb E0q þ x0d Vb cos d  ðx0d þ xtb Þvesq
 xes iesq þ ðxbs þ xbt Þibsq ¼ vbsd  vesd þ Vb sin d
xes iesd  ðxbs þ xbt Þibsd ¼ vbsq  vesq þ Vb cos d

which in the matrix form is


   1
iesq xq xtb þ ðxte þ xes Þðxq þ xtb Þ xq xtb þ xte ðxq þ xtb Þ
¼
ibsq xes xes þ xbt
 
ðxq þ xtb Þme ke Vdc cos wes  xq Vb sin d

mb kb Vdc cos wbs  me ke Vdc cos wes þ Vb sin d
    ð4:82Þ
iesd ðxte þ xes Þðx0d þ xtb Þ þ x0d xte xte ðx0d þ xtb Þ þ x0d xtb 1
¼
ibsd xes ðxbs þ xbt Þ
 
xtb E0q þ x0d Vb cos d  ðx0d þ xtb Þme ke Vdc sin wes

mb kb Vdc sin wbs  me ke Vdc sin wes þ Vb cos d
148 4 Single-Machine Infinite-Bus Power Systems …

For the single-machine infinite-bus power system with the UPFC installed, it can
have

ited ¼ iesd þ ibsd


iteq ¼ iesq þ ibsq
Pt ¼ vtq ðiteq þ itbq Þ þ vtd ðited þ itbd Þ ð4:83Þ
Eq ¼ E0q þ ðxd  x0d Þðited þ itbd Þ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Vt ¼ v2td þ v2tq

Equations (4.74), (4.75), (4.78), (4.80), (4.82), and (4.83), together with the dif-
ferential equations of synchronous generator of Eq. (2.35), form the nonlinear
mathematical model of the single-machine infinite-bus power system with the
UPFC installed.

4.3.1.3 Linearized Model of the Single-Machine Infinite-Bus Power


System Installed with the UPFC

Linearization of Eq. (4.82) is


2 3 2 3
2 3 Dme  2 3 Dme
  Dd  Dd
Diesq 6 Dwes 7 Diesd 6 7
0 6 Dwes 7
¼ F0q 4 DE0q 5 þ G0q 6 7 04 0 5
4 Dmb 5; Dibsd ¼ Fd DEq þ Gd 4 Dmb 5
Dibsq
DVdc DVdc
Dwbs Dwbs
ð4:84Þ

where
  
1 xbs þ xbt xte ð2xq þ xtb Þ f q11 0 f q13
F0q ¼
xR1 xes xq xte þ ðxte þ xes Þðxq þ xtb Þ f q21 0 f q23
  
0 1 xbs  xbt xte xtb f d11 xL f d13
Fd ¼
xR2 xes ðxte þ xes Þðx0d þ xtb Þ  x0d xte f d21 0 f d23
 " #
1 x bs þ x bt x te ð2x q þ x tb Þ g g 0 0
G0q ¼
q11 q12
xR1 xes xq xte þ ðxte þ xes Þðxq þ xtb Þ gq21 gq22 gq23 gq24
  
0 1 xbs  xbt xte xtb gd11 gd12 0 0
Gd ¼
xR2 xes ðxte þ xes Þðx0d þ xtb Þ  x0d xte gd21 gd22 gd23 gd24
xR1 ¼ ½xq xte þ ðxte þ xes Þðxq þ xtb Þðxbs þ xbt Þ þ x2es ð2xq þ xtb Þ
xR2 ¼ ½ðxte þ xes Þðx0d þ xtb Þ  x0d xte ðxbs þ xbt Þ  x2te xtb
4.3 Damping Function of a Unified Power Flow Controller … 149

f q11 ¼ ðxq þ xtb Þme0 ke Vdc0 sin wes0  xq Vb cos d0


f q13 ¼ ðxq þ xtb Þme0 ke cos wes0
f q21 ¼ mb0 kb Vdc0 sin wbs0  me0 ke Vdc0 sin wes0 þ Vb cos d0
f q23 ¼ mb0 kb cos wbs0  me0 ke cos wes0
f d11 ¼ ðx0d þ xtb Þme0 ke Vdc0 cos wes0  x0d Vb sin d0
f d13 ¼ ðx0d þ xtb Þme0 ke sin wes0
f d21 ¼ me0 ke Vdc cos wes0  mb0 kb Vdc0 cos wbs0  Vb sin d0
f d23 ¼ mb0 kb sin wbs0  me0 ke sin wes0
gq11 ¼ ðxq þ xtb Þke Vdc0 cos wes0
gd12 ¼ ðxq þ xtb Þme0 kE Vdc0 sin wes0
gq21 ¼ ke Vdc0 cos wes0
gq22 ¼ me0 ke Vdc0 sin wes0
gq23 ¼ kb Vdc0 cos wbs0
gq24 ¼ mb0 kb Vdc0 sin wbs0
gd11 ¼ ðx0d þ xtb Þke Vdc0 sin wes0
gd12 ¼ ðx0d þ xtb Þme0 ke Vdc0 cos wes0
gd21 ¼ ke Vdc0 sin we0
gd22 ¼ me0 ke Vdc cos wes0
gd23 ¼ kb Vdc0 sin wbs0
gd24 ¼ mb0 kb Vdc0 cos wbs0

From Eqs. (4.66) and (4.67), it can have

vetd þ jvetq ¼ jxes ðiesd þ jiesq Þ þ me ke Vdc cos wes þ jme ke Vdc sin wes
vbtd þ jvbtq ¼ jxbs ðibsd þ jibsq Þ þ mb kb Vdc cos wbs þ jmb kb Vdc sin wbs

The above equations give

vesd ¼ xes iesq þ me ke Vdc cos wes


vesq ¼ xes iesd þ me ke Vdc sin wes
vbsd ¼ xbs ibsq þ mb kb Vdc cos wbs
vbsq ¼ xbs ibsd þ mb kb Vdc sin wbs
150 4 Single-Machine Infinite-Bus Power Systems …

By using Eq. (4.84), linearization of the above equation can be obtained as

Dvesd ¼ xes f 0q11 Dd  xE f 0q12 DE0q  ðxes f 0q13  me0 ke cos wes0 ÞDVdc
 ðxes g0q11  ke Vdc0 cos wes0 ÞDme  ðxes g0q12 þ me0 ke Vdc0 sin wes0 ÞDwes
 xes g0q13 Dmb  xes g0q14 Dwbs
¼ aed1 Dd þ aed2 DE0q þ aed3 DVdc þ aed4 Dme þ aed5 Dwes þ aed6 Dmb þ aed7 Dwbs
Dvesq ¼ xes f 0d11 Dd þ xes f 0d12 DE0q þ ðxes f 0d13 þ me0 ke sin wes0 ÞDVdc
þ ðxes g0d11 þ ke Vdc0 sin wes0 ÞDme
þ ðxes g0d12 þ me0 ke Vdc0 cos wes0 ÞDwes þ xes g0d13 Dmb þ xes g0d14 Dwbs
¼ aeq1 Dd þ aeq2 DE0q þ aeq3 DVdc þ aeq4 Dme þ aeq5 Dwes þ aeq6 Dmb þ aeq7 Dwbs
Dvbsd ¼ xbs f 0q21 Dd  f 0q22 DE0q  ðxbs f 0q23  mb0 kb cos wbs0 ÞDVdc
 ðxbs g0q23  kb Vdc0 cos wbs0 ÞDmb  ðxbs g0q24 þ mb0 kb Vdc0 sin wbs0 ÞDwbs
 xbs g0q21 Dme  xbs g0q22 Dwes
¼ abd1 Dd þ abd2 DE0q þ abd3 DVdc þ abd4 Dme þ abd5 Dwes þ abd6 Dmb þ abd7 Dwbs
Dvbsq ¼ xbs f 0d21 Dd þ xbs f 0d22 DE0q þ ðxbs f 0d23 þ mb0 kb sin wbs0 ÞDVdc
þ ðxbs g0d23 þ kb Vdc0 sin wbs0 ÞDmb þ ðxbs g0d24 þ mb0 kb Vdc0 cos wbs0 ÞDwbs
þ xbs g0d21 Dme þ xbs g0d22 Dwes
¼ abq1 Dd þ abq2 DE0q þ abq3 DVdc þ abq4 Dme þ abq5 Dwes þ abq6 Dmb þ abq7 Dwbs

By using the above equations, linearization of Eq. (4.75) can be obtained as

Dwes ¼ a0e1 Dd þ a0e2 DE0q þ a0e3 DVdc þ a0e4 Dme


þ a0e5 Dwes þ a0e6 Dmb þ a0e7 Dwbs  Dde
Dwbs ¼ a0b1 Dd þ a0b2 DE0q þ a0b3 DVdc þ a0b4 Dme
þ a0b5 Dwes þ a0b6 Dmb þ a0b7 Dwbs  Ddb

where

1
a0ei ¼ ðvesd0 aeqi  vesq0 aedi Þ
v2esd0 þ v2esq0
1
a0bi ¼ 2 ðvbsd0 abqi  vbsq0 abdi Þ
vbsd0 þ v2bsq0
i ¼ 1; 2; 3; 4; 5; 6; 7
4.3 Damping Function of a Unified Power Flow Controller … 151

Rearrangement of above equations is


2 3
   1   Dd
Dwes 1  a0e5 a0e7 a0e1 a0e2a0e3 6 7
¼ 4 DE0q 5
Dwbs a0b5 1  a0b7 a0b1 a0b2a0b3
DVdc
     !
a0e4 a0e6 Dme 1 0 Dde
þ þ
a0b4 a0b6 Dmb 0 1 Ddb
2 3
Dd    
6 0 7
Dme Dde
¼ AEB1 4 DEq 5 þ AEB2 þ AEB3 ð4:85Þ
Dmb Ddb
DVdc

Substituting Eq. (4.85) into Eq. (4.84), it can have


3 2
2 3
Dme
  Dd 6 Dd 7
Diesq 6 7 6 e7
¼ Fq 4 DE0q 5 þ Gq 6 7
Dibsq 4 Dmb 5
DVdc
Ddb
2 3 ð4:86Þ
2 3 Dme
  Dd 6 Dd 7
Diesd 6 7 6 e7
¼ Fd 4 DE0q 5 þ Gd 6 7
Dibsd 4 Dmb 5
DVdc
Ddb

Linearization of Eq. (4.80) is


  
xq Vb sin d0 xq xtb xq xtb Diesq
Dvtd ¼ Dd þ
xq þ xtb xq þ xtb
xq þ xtb Dibsq
   ð4:87Þ
0
xd Vb sin d0 xtb 0 x0d xtb x0d xtb Diesd
Dvtq ¼ Dd þ 0 DE þ 0
x0d þ xtb xd þ xtb q xd þ xtb x0d þ xtb Dibsd

Substituting Eq. (4.86) into (4.87) gives


2 3
2 3 Dme
Dd 6 Dd 7
6 7 6 e7
Dvtd ¼ Hd 4 DE0q 5 þ Ld 6 7
4 Dmb 5
DVdc
Ddb
2 3 ð4:88Þ
2 3 Dme
Dd 6 Dd 7
6 7 6 e7
Dvtq ¼ Hq 4 DE0q 5 þ Lq 6 7
4 Dmb 5
DVdc
Ddb
152 4 Single-Machine Infinite-Bus Power Systems …

where
   
xq Vb sin d0 xq xtb xq xtb
Hd ¼ 0 0 þ F
xq þ xtb xq þ xtb xq þ xtb q
 
xq xtb xq xtb
Ld ¼ Gq
xq þ xtb xq þ xtb
 0   
xd Vb sin d0 xtb x0d xtb x0d xtb
Hq ¼ 0 þ 0 Fd
x0d þ xtb x0d þ xtb xd þ xtb x0d þ xtb
 
x0 xtb x0d xtb
Lq ¼ 0 d Gd
xd þ xtb x0d þ xtb

Linearization of Eq. (4.78) is


Vb sin d0 1
Ditbd ¼ Dd þ Dvtq
xtb xtb
ð4:89Þ
Vb cos d0 1
Ditbq ¼ Dd  Dvtd
xtb xtb

Linearization of Eq. (4.83) is

Dited ¼ Diesd þ Dibsd


Diteq ¼ Diesq þ Dibsq
DPt ¼ ðiteq0 þ itbq0 ÞDvtq þ ðited0 þ itbd0 ÞDvtd
þ vtq0 ðDiteq þ Ditbq Þ þ vtd0 ðDited þ Ditbd Þ ð4:90Þ
DEq ¼ DE0q þ ðxd  x0d ÞðDited þ Ditbd Þ
vtd0 vtq0
DVt ¼ Dvtd þ Dvtq
Vt0 Vt0

By using Eqs. (4.86), (4.88), and (4.89) and the first two equations in Eq. (4.90), the
last three equations in Eq. (4.90) become
DPt ¼ ðiteq0 þ itbq0 ÞDvtq þ ðited0 þ itbd0 ÞDvtd
þ vtq0 ðDiesq þ Dibsq þ Ditbq Þ þ vtd0 ðDiesd þ Dibsd þ Dibsd Þ

Vb cos d0 Vb sin d0
¼ vtq0 þ vtd0 Dd
xtb xtb

vtq0 vtd0
þ ited0 þ itbd0  Dvtd þ iteq0 þ itbq0 þ Dvtq
xtb xtb
   
Diesq Diesd
þ ½ vtq0 vtq0  þ ½ vtd0 vtd0 
Diesq Diesd
2 3
2 3 Dme
Dd 6 Dd 7
6 7 6 e7
¼ ½ K1 K2 Kpdc 4 DE0q 5 þ ½ Kpme Kpde Kpmb Kpdb 6 7 ð4:91Þ
4 Dmb 5
DVdc
Ddb
4.3 Damping Function of a Unified Power Flow Controller … 153

 
Diesd
DEq ¼ DE0q
þ ½ ðxd  x0d Þ
ðxd  x0d Þ 
Dibsd

0 Vb sin d 0 1
þ ðxd  xd Þ Dd þ Dvtq
xtb xtb
Vb sin d0 1
¼ ðxd  x0d Þ Dd þ DE0q þ ðxd  x0d Þ Dvtq
xtb xtb
 
0 0
Diesd
þ ½ ðxd  xd Þ ðxd  xd Þ 
Dibsd
2 3
2 3 Dme
Dd 6 Dd 7
6 7 6 e7
¼ ½ K4 K3 Kqdc 4 DE0q 5 þ ½ Kqme Kqde Kqme Kqde 6 7
4 Dmb 5
DVdc
Ddb
ð4:92Þ
2 3
2 3 Dme
Dd 6 Dde 7
DVt ¼ ½ K5 K6 Kvdc 4 DE0q 5 þ ½ Kvme Kvde Kvmb Kvdb 6 7
4 Dmb 5 ð4:93Þ
DVdc
Ddb

where
h i
½ K1 K2 Kpdc  ¼ vtq0 Vbxcos d2 þ vted0 Vbxsin d0 0 0

tb tb

vteq0 vted0
þ ited0 þ itbd0  Hd þ iteq0 þ itbq0 þ Hq
xtb xtb

þ ½ vtq0 vtq0 Fq þ ½ vtd0 vtd0 Fd

vteq0
½ Kpme Kpde Kpmb Kpdb  ¼ ðited0 þ itbd0  ÞLd
xtb
vted0
þ ðiteq0 þ itbq0 þ ÞLq þ ½ vtq0 vtq0 Gq þ ½ vtd0 vtd0 Gd
xtb
h i
½ K4 K3 Kqdc  ¼ ðxd  x0d Þ Vbxsin d0 1 0
tb

ðxd  x0d Þ
þ Hq þ ½ ðxd  x0d Þ ðxd  x0d Þ Fd
xtb
ðxd  x0d Þ
½ Kqme Kqde Kqmb Kqdb  ¼ Lq þ ½ ðxd  x0d Þ ðxd  x0d Þ Gd
xtb
vtd0 vtq0
½ K5 K6 Kvdc  ¼ Hd þ Hq
Vt0 E
vtd0 vtq0
½ Kqme Kqde Kqmb Kqdb  ¼ Ld þ Lq
Vt0 Vt0
154 4 Single-Machine Infinite-Bus Power Systems …

Linearization of Eq. (4.74) is


 
Diesd
Cdc DV_ dc ¼ ½ me0 ke cos wes0 mb0 kb cos wbs0 
Dibsd
 
Diesq
þ ½ me0 ke sin wes0 mb0 kb sin wbs0 
Dibsq
þ ke iesd0 cos wes0 Dme þ ke iesq0 sin wes0 Dme þ kb ibsd0 cos wbs0 Dmb
þ kb ibsq0 sin wbs0 Dmb  me0 ke iesd0 sin wes0 ðDd þ Dde Þ
þ me0 ke iesq0 cos wes0 ðDd þ Dde Þ  mb0 kb ibsd0 sin wbs0 ðDd þ Ddb Þ
þ mb0 kb ibsq0 cos wbs0 ðDd þ Ddb Þ
¼ ðme0 ke iesd0 sin wes0  me0 ke iesq0 cos wes0 þ mb0 kb ibsd0 sin wbs0
 
Diesd
 mb0 kb ibsq0 cos wbs0 ÞDd þ ½ me0 ke cos wes0 mb0 kb cos wbs0 
DiBbsd
 
Diesq
þ ½ me0 ke sin wes0 mb0 kb sin wbs0 
Dibsq
þ ðke iesd0 cos wes0 þ ke iesq0 sin wes0 ÞDme
þ ðkb ibsd0 cos wbs0 þ kb ibsq0 sin wbs0 ÞDmb
þ ðme0 ke iesq0 cos wes0  me0 ke iesd0 sin wes0 ÞDde
þ ðmb0 kb ibsq0 cos wbs0  mb0 kb ibsd0 sin wbs0 Ddb ÞDdb
2 3
2 3 Dme
Dd 6 Dd 7
6 7 6 e7
¼ ½ K7 K8 K9 4 DE0q 5 þ ½ Kdme Kdde Kdmb Kddb 6 7
4 Dmb 5
DVdc
Ddb
ð4:94Þ

where
½ K7 K8 K9 
2 3T
me0 ke iesd0 sin wes0  me0 ke iesq0 cos wes0 þ mb0 kb ibsd0 sin wbs0  mb0 kb ibsq0 cos wbs0
6 7
¼4 0 5
0

þ ½ me0 ke cos wes0 mb0 kb cos wB0 Fd þ ½ me0 ke sin wE0 mb0 kb sin wB0 Fq

2 3T
ke iesd0 cos wes0 þ ke iesq0 sin wes0
6 me0 ke iesq0 cos w  me0 ke iesd0 sin w 7
6 es0 es0 7
½ Kdme Kdde Kdmb Kddb  ¼ 6 7
4 kb ibsd0 cos wbs0 þ kb ibsq0 sin wbs0 5
mb0 kb ibsq0 cos wbs0  mb0 kb ibsd0 sin wbs0

þ ½ me0 ke cos wes0 mb0 kb cos wbs0 Gd þ ½ me0 ke sin wes0 mb0 kb sin wbs0 Gq
4.3 Damping Function of a Unified Power Flow Controller … 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 Heffron–Phillips model of the single-machine infinite-bus power system
with the UPFC installed can be obtained as
2 3
2 3 0 xo 0 0 0 2 3
Dd_ 6
6  K1 D K2 Kpdc 7 7 Dd
6 _ 7 6 M  M  M 0  M 76 Dx 7
6 Dx 7 6 Kqdc 7 6 7
6 7 6 K K  0 7 6 7
6 DE_ 0 7 ¼ 6  T0  0 1
0 7 6 0 7
4 3

6 q 7 6 d0 Td0 0
Td0 Td0 76 DE q 7
6 7 6 76 7
4 DE_ 0fd 5 6  KA K5 0  KTA K6  T1  KATKvdc 7 4 DE0fd 5
6 T 7
DV_ dc 4 A A A A
5 DVdc
K7 0 K8 0 K9
Cdc Cdc Cdc ð4:95Þ
2 3
0 0 0 0
6  Kpme K
 Mpde
Kpmb
 M  Mpdb 7
K 2 3
6 M 7 Dme
6 7
6 K K K K 76 Dd 7
6  qme  qde  qmB  qdB 76 e 7
6 T0d0 T0d0 T0d0 T0d0 76 7
6 74 5
6  KA Kvme  KA Kvde  KA Kvmb  KA Kvdb 7 Dmb
6 7
TA 5 Ddb
4 TA TA TA
Kdme Kdde Kdmb Kddb
Cdc Cdc Cdc Cdc

The linearized model is shown in Figs. 4.14 and 4.15, where

kp ¼ ½ Kpdc Kpme Kpde Kpmb Kpdb 


kq ¼ ½ Kqdc Kqme Kqde Kqmb Kqdb 
kv ¼ ½ Kvdc Kvme Kvde Kvmb Kvdb 
ð4:96Þ
kd ¼ ½ Kdme Kdde Kdmb Kddb 
uupfc1 ¼ ½ Dme Dde Dmb Ddb T
 T
uupfc ¼ DVdc uTupfc1

4.3.2 Design of a UPFC Stabilizer Installed


in a Single-Machine Infinite-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 4 Single-Machine Infinite-Bus Power Systems …

K1

1 Δω ω0 Δδ
- sM + D s

K4 K5
K2
kp uupfc kv

kq

1 KA
ΔE q ' Td0 's + K 3 - + 1 + sTA -
K6

Fig. 4.14 Extended Heffron–Phillips model of power system with a UPFC installed—part of the
power system

Fig. 4.15 Extended Heffron– Δδ


Phillips model of power
system with a UPFC installed
—part of the UPFC K7

1
ΔE q ' K8 + ΔVdc
sCdc − K 9

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 infinite-bus power system installed with
the UPFC of Eq. (4.95) can be written as
4.3 Damping Function of a Unified Power Flow Controller … 157

X
4
X_ ¼ AX þ bk Duk ð4:97Þ
k¼1

where
2 3
2 3 0 xo 0 0 0
Dd 6 K 7
6  1K MD K 2
0  Mpdc 7
6 7 6 M M 7
6 Dx 7 6 Kqdc 7
6 7 6  K0 4  K0 3 1
 0 7
X¼6 0 7
6 DEq 7; A ¼ 6
6 Td0
0
Td0 T0d0 Td0 7 7;
6 0 7 6 K K 7
4 DEfd 5 6 A 5 0  TA K6
K T  T 7
1 K A Kvdc 7
6 TA
DVdc 4 A A A
5
K7 0 K8 0 K9
Cdc Cdc Cdc
2 3 2 3 2 3 2 3
0 0 0 0
6  Kpme 7 6  Kpde 7 6  Kpmb 7 6  Kpdb 7
6 M 7 6 M 7 6 M 7 6 M 7
6 7 6 7 6 7 6 7
6 Kqme 7 6 Kqde 7 6 Kqmb 7 6 Kqdb 7
6  0 7 6  0 7 6  0 7 6  0 7
b1 ¼ 6 Td0 7; b2 ¼ 6 Td0 7; b3 ¼6 Td0 7; b4 ¼ 6 Td0 7;
6 7 6 7 6 7 6 7
6  KA Kvme 7 6  KA Kvde 7 6  KA Kvmb 7 6  KA Kvdb 7
6 TA 5 7 6 TA 7 6 TA 5 7 6 TA 7
4 4 5 4 4 5
Kdme Kdde Kdmb Kddb
Cdc Cdc Cdc Cdc
Du1 ¼ Dme ; Du2 ¼ Dde ; Du3 ¼ Dmb ; Du4 ¼ Ddb

If the transfer function and feedback signal of the UPFC stabilizer are Tupfck ðs) and
yk , respectively, state-space representation of the power system with the UPFC
stabilizer is

X
4
X_ ¼ AX þ bk Duk
k¼1
ð4:98Þ
yk ¼ ckT X
Duk ¼ Tupfck ðsÞyk

If ki is the electromechanical oscillation mode of the power system, from Eqs. (2.50
) and (2.54), it can have

Avi ¼ ki vi
ð4:99Þ
wTi A ¼ wTi ki

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 4 Single-Machine Infinite-Bus Power Systems …

X
4 X
4
Z_ ¼ V1 AVZ þ V1 bk Duk ¼ K Z þ V1 bk Duk
k¼1 k¼1
ð4:100Þ
yk ¼ cTk VZ
Duk ¼ Tupfck ðsÞyk

where
2 3
k1 0 0 0
60 k2 0 07
6 7
V ¼ ½ v1 v2 . . . vn ; K¼6 .. 7
40 0 . 05
0 0 0 kn

According to Eq. (2.53),

V1 ¼ ½ w1 w2 . . . w M T ¼ W T ð4:101Þ

Equation (4.100) can be written as

1 X 4
zj ¼ þ wTj bk Duk
s  kj k¼1
X
yk ¼ cTk vj zj ; j ¼ 1; 2; . . . ð4:102Þ
j

Duk ¼ Tupfck ðsÞyk

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 defined that as

Fig. 4.16 Modal 1


decomposition of the system b k1 Ck1
to be installed with UPFC s − λ1
stabilizer
1
bk 2 Ck 2 +
s − λ2

1
b kn Ckn
s − λn

Δu k yk
Tupfck (s)
4.3 Damping Function of a Unified Power Flow Controller … 159

far as the oscillation mode ki is concerned, the controllability index and observ-
ability index of the UPFC stabilizer are, respectively,

bik ¼ wiT bk ; cik ¼ ckT vi ð4:103Þ

Their product is the residue

Rik ¼ bik cik ð4:104Þ

which measures the overall influence of the stabilizer on the oscillation mode.
Obviously, the controllability index, observability index, and residue are cal-
culated from the open-loop system model as far as the UPFC stabilizer is con-
cerned. 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 modu-
lation 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 ¼ Dme ; Du2 ¼ Dde ; Du3 ¼ Dmb ; Du4 ¼ Ddb


when k ¼ 1; kpuk ¼ Kdme ;
kpuk ¼ ½ Kpdc Kpme ; kquk ¼ ½ Kqdc Kqme ; kvuk ¼ ½ Kvdc Kvme 
when k ¼ 2; kpuk ¼ Kdde ;
kpuk ¼ ½ Kpdc Kpde ; kquk ¼ ½ Kqdc Kqde ; kvuk ¼ ½ Kvdc Kvde 
when k ¼ 3; kpuk ¼ Kdmb ;
kpuk ¼ ½ Kpdc Kpmb ; kquk ¼ ½ Kqdc Kqmb ; kvuk ¼ ½ Kvdc Kvmb 
when k ¼ 4; kpuk ¼ Kddb ;
kpuk ¼ ½ Kpdc Kpdb ; kquk ¼ ½ Kqdc Kqdb ; kvuk ¼ ½ Kvdc Kvdb 

The forward path is described by the following state-space representation

X_ 1 ¼ A1 X1 þ b1k Duk
ð4:105Þ
DTuk ¼ cTupfc1 X1 þ kpuk Duk ; k ¼ 1; 2; 3; 4
160 4 Single-Machine Infinite-Bus Power Systems …

Fig. 4.17 Forward path of ΔTuk


UPFC stabilizer K8 + K duk Δu k

- 1
sCdc − K 9

K2
⎡ ΔVdc ⎤
k puk ⎢ Δu ⎥
⎣ k⎦ k vuk

k quk

1 KA
ΔE q ' Td0 's + K 3 - + 1 + sTA -
K6

where
Du1 ¼ Dme ; Du2 ¼ Dde ; Du3 ¼ Dmb ; Du4 ¼ Ddb
2 3
2 3 K
0  K0 3 1
0  qdc0
DEq 6 Td0 Td0 Td0 7
6 7 6 7
X1 ¼ 4 DE0fd 5; A1 ¼ 6 K
6  TA
A K6 K
T  T 7
1 A Kvdc 7

4 A A 5
DVdc K8 K
Cdc 0 9
Cdc
T T
b11 ¼ ½ Kqme Kvme Kdme  ; b12 ¼ ½ Kqde Kvde Kdde  ;
b13 ¼ ½ Kqmb Kvmb Kdmb T ; b14 ¼ ½ Kqdb Kvdb Kddb T ;
cupfc1 ¼ ½ K2 0 Kpdc T

From Eq. (4.105), transfer function of the forward path can be obtained as

Fupfck ðsÞ ¼ cupfc1


T
ðsI  A1 Þ1 b1k ; k ¼ 1; 2; 3; 4 ð4:106Þ

Hence, the electric torque contributed from the UPFC stabilizer is


DTuk ¼ Fupfck ðjxs ÞTupfck ðjxs Þyk ; k ¼ 1; 2; 3; 4 ð4:107Þ

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 ðjxs Þ ; k ¼ 1; 2; 3; 4 ð4:108Þ
4.3 Damping Function of a Unified Power Flow Controller … 161

4.3.2.3 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 configurations). Design of the stabilizer is carried out based on the lin-
earized 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 ¼ Maxðbik ¼ uTi bk Þ; Duk 2 fDme ; Dmb ; Dde ; Ddb g ð4:109Þ


Duk

When the damping torque index is used, according to Eq. (4.108), the criterion of
selection is
 
Duselected ¼ Max Fupfck ðjxs Þ ; k ¼ 1; 2; 3; 4 ; Duk 2 fDme ; Dmb ; Dde ; Ddb g
Duk

ð4:110Þ

To consider the robustness of stabilizer to the variations of system operating con-


ditions, the above criteria of Eqs. (4.109) and (4.110) should be calculated at a
selected operating condition of the power system when

rselected ¼ Minðbik ¼ uTi bk Þ; r 2 X0 ð4:111Þ


r
162 4 Single-Machine Infinite-Bus Power Systems …

or
 
rselected ¼ Min Fupfck ðjxs Þ ; k ¼ 1; 2; 3; 4 ; r 2 X0 ð4:112Þ
r

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 infinite-bus
power system to be Tstab ðsÞ ¼ Kstab Tstabk ðsÞ, 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 ¼ Kdstab Dx þ Ksstab Dd ¼ Kdstab þ Dx ¼ Ky ðjxs ÞDx
jxs
ð4:113Þ

If the forward path of stabilizer to the electromechanical oscillation loop of gen-


erator is Fstab ðs), the damping torque provided by the stabilizer is

DTDstab ¼ Re½Fstab ðjxs ÞKy ðjxs ÞTstab ðjxs ÞDx ð4:114Þ

and the synchronizing torque is


xs
DTSstab ¼ Im½Fstab ðjxs ÞKy ðjxs ÞTstab ðjxs Þ Dd ð4:115Þ
x0

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 ðjxs ÞKy ðjxs Þ. At an operating condition rj 2 X0 , if
it is denoted that

Fstab ðjxs ÞKy ðjxs Þ ¼ Hj \uj ; at rj 2 X0 ð4:116Þ

Tstab ðjxs Þ ¼ H\  / ð4:117Þ

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 ¼ Hj H cosðuj  /ÞDx ð4:118Þ

xs
DTsstab ¼ Hj H sinðuj  /Þ Dd ð4:119Þ
x0
4.3 Damping Function of a Unified Power Flow Controller … 163

If it has

umax ¼ maxðuj Þ; rj 2 X0 ð4:120Þ

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 ð4:120aÞ

Hence, simply it can choose


/ ¼ umax ð4:121Þ

Therefore, from Eqs. (4.118) and (4.121), the damping torque contributed by the
stabilizer can be obtained as

DTdstab ¼ Hj H cosðuj  umax ÞDx ð4:122Þ

Equation (4.122) indicates that the effectiveness of stabilizer to be designed is


measured by Hj cosðuj  umax Þ 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 cosðuselected  umax Þ ¼ Min½Hj cosðuj  umax Þ; rj 2 X0  ð4:123Þ


j

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 Infinite-Bus Power


System Installed with a BESS Stabilizer

4.4.1.1 Extended Heffron–Phillips Model of the Example Power


System Installed with a BESS

Configuration of an example single-machine infinite-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 4 Single-Machine Infinite-Bus Power Systems …

rbess
Vc

VSC Cdc Vdc Vbess

BESS

Fig. 4.18 Model of BESS

Vbess  Vdc
Idc2 ¼ ð4:124Þ
rbess

Parameters of the example system are as follows:


Generator xd ¼ 1:0; xq ¼ 0:6; x0d ¼ 0:3; M ¼ 8s:; D ¼ 0; T0d0 ¼ 5:044s:
The AVR KA ¼ 100; TA ¼ 0:01
Transmission line xts ¼ 0:3; xsb ¼ 0:3; xs ¼ 0:15
The BESS k ¼ 3; Cdc ¼ 1:0; Vdc0 ¼ 1:0; Vbess ¼ 1:0; rbess ¼ 0:01
Steady-state oper- Pt0 ¼ 0:5; Vt0 ¼ 1:0; Vb ¼ 1:0; Vs0 ¼ 1:0
ating point

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:
2sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3

2
Vb 4 x P
 Vb 5 ¼ 0:0337
sb t0
Qb0 ¼ V2s0 
Xsb Vb
Pt0  jQb0
Isb0 ¼ ¼ 0:5 þ j0:0377
Vb
Vs0 ¼ jxsb Isb þ Vb ¼ 0:9887 þ j0:15 ¼ 1:0\8:6
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
V2t0 V2s0  ðPt0 xts Þ2  V2s0
Qs0 ¼ ¼ 0:0377
xts
Pt0  jQs0
Its0 ¼ ¼ 0:4887 þ j0:1123
Vs0
Is0 ¼ Its0  Isb0 ¼ 0:0113 þ j0:0746
Vt0 ¼ jxts Its0 þ Vs0 ¼ 0:955 þ j0:2966
EQ ¼ Vt0 þ jxq Its0 ¼ 0:8876 þ j0:5898 ¼ 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 ¼ ReðIts0 Þ sin d0  ImðIts0 Þ cos d0 ¼ 0:1769


itsq0 ¼ ReðIts0 Þ cos d0 þ ImðIts0 Þ sin d0 ¼ 0:4692
vtd0 ¼ ReðVt0 Þ sin d0  ImðVt0 Þ cos d0 ¼ 0:2815
vtq0 ¼ ReðVt0 Þ cos d0 þ ImðVt0 Þ sin d0 ¼ 0:9596
isd0 ¼ ReðIs0 Þ sin d0  ImðIs0 Þ cos d0 ¼ 0:0684
isq0 ¼ ReðIs0 Þ cos d0 þ ImðIs0 Þ sin d0 ¼ 0:0318

E0q0 ¼ EQ  ðxd  x0d Þitsd0 ¼ 1:0126


Vc0 ¼ Vs0  jxs Is0 ¼ 0:9999 þ j0:1517 ¼ 1:011\8:6
Vc0
m0 ¼ ¼ 0:3371
kVdc0

From Fig. 4.3, it can be seen that w0 and wx0 ¼ 8:6 is the phase of Vc0 in the d–
q coordinate and x–y coordinate, respectively. Hence,

p
w0 ¼ wx0 þ  d0 ¼ 65:0250
2

From Eqs. (4.7) and (4.15), it is obtained that

       
c11 c12 0:6667 1:0000 d11 d12 0:9524 1:4286
¼ ; ¼
c21 c22 2:6667 0:6667 d21 d22 2:8571 0:9524

By using Eqs. (4.16)–(4.19), following coefficients in the Heffron–Phillips


model of Figs. 4.4 and 4.5 are obtained

K1 ¼ 0:3330; K2 ¼ 0:6702; K3 ¼ 2; K4 ¼ 0:1845; K5 ¼ 0:0290;


K6 ¼ 0:5483; K7 ¼ 1:9772; K8 ¼ 0:4067; K9 ¼ 0:0746
Kdm ¼ 0:2212; Kpm ¼ 0:5355; Kqm ¼ 1:8130; Kpw ¼ 0:7086;
Kqw ¼ 0:2847; Kdw ¼ 2:6858; Kpdc ¼ 0:1805; Kqdc ¼ 0:6112;
Kvm ¼ 0:8882; Kvw ¼ 0:0138; Kvdc ¼ 0:2994
166 4 Single-Machine Infinite-Bus Power Systems …

From Eq. (4.124), it can have

1
DIdc2 ¼  DVdc ð4:125Þ
rbess

From Eqs. (4.19) and (4.125), state equation of the example power system is
obtained as

X_ ¼ AX þ Bu ð4:126Þ

where

2 3
Dd
6 7
6 Dx 7  
6 7 D/
6 0 7
X ¼ 6 DEq 7; u ¼ ;
6 7 Dm
4 DE0fd 5
DVdc
2 3
0 xo 0 0 0
6 K 7
6 K 1
MD K 2
0  Mpdc 7
6 M M 7
6 Kqdc 7
6  K0 4 0  K3 1 7
A¼6 T T T 0
T 0
7
6 d0 d0 d0 d0 7
6 KA K5 K A K6 K 7
6 0  T T 1
 T vdc 7
4 TA A A A 5
K7 0 K8 0 K9  C 1r
2 dc bess
3
0 314:16 0 0 0
6 7
6 0:1302 0 0:0838 0 0:0226 7
6 7
¼66 0:0930 0 0:3965 0:1983 0:1212 7;
7
6 7
4 428:242 0 5483:2 100 2994:2 5
0:7086 0 0:4067 0 100:0746
2 3
0 0 2 3
6  Kpw Kpm 7 0 0
6 M  M 7 6 7
6 7 6 0:0886 0:0669 7
6 Kqw K 7 6 7
B¼6
6
 0
Td0
 qm 7 6
T0d0 7 ¼ 6 0:0564 0:3594 7
7
6 7 6 7
6 KA Kvw 7 4 138:495 8882 5
4 T  KATKvm 5
A A
2:6858 0:2212
Kdw Kdm
4.4 Examples 167

Eigenvalues of state matrix are


k1 ; k2 ¼ 0:0591  j6:4978
k3 ¼ 13:0310
k4 ¼ 100:2902
k5 ¼ 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 Heffron–Phillips
model of Figs. 4.4 and 4.5 can be ignored. Hence, from Fig. 4.4 with Dd being
ignored, it can have

1
DVdc ¼ ðK8 DE0q þ Kdm Dm þ Kdw D/Þ ð4:127Þ
s  K9 þ C 1r
dc bess

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
DE0q ¼ ðKvdc DVdc  Kvw Dw  Kvm Dm  K6 DE0q Þ
sT0d0 þ K3 1 þ sTA

ð4:128Þ
Kqdc DVdc  Kqw D/  Kqm Dm

DTet ¼ K2 DE0q þ Kpdc DVdc þ Kpw D/ þ Kpm Dm

-
ΔTet
[ ΔVdc Δψ Δm ]
K2

⎣⎡ K pdc K pψ K pm ⎦⎤ ⎣⎡ K vdc K vψ K vm ⎦⎤

⎣⎡ K qdc K qm K qψ ⎦⎤

ΔE q '
-
1 KA
K 3 + sTd0 ' + 1 + sTA -
K6

Fig. 4.19 Forward path of BESS


168 4 Single-Machine Infinite-Bus Power Systems …

Deleting variable DVdc and DE0q in Eqs. (4.126) and (4.127), it can be obtained
that

DTes ¼ Fact ðsÞD/ þ Frct ðsÞDm ð4:129Þ

With xs ¼ Imðk1 or k2 Þ ¼ 6:4978, it can be computed that

Fact ðjxs Þ ¼ 0:7164  j0:0017 ¼ 0:7164\  0:136


ð4:130Þ
Frct ðjxs Þ ¼ 0:2467 þ j0:4706 ¼ 0:5313\117:661

A simple way to calculate the transfer function of the forward path with s ¼ jxs
is to obtain the following state-space representation of the system of Eqs. (4.127)
and (4.128)

X_ r ¼ Ar Xr þ Br u
ð4:131Þ
DTet ¼ Cr Xr þ Dr u
2 3
DE0q  
D/
where Xr ¼ 4 DE0fd 5; u ¼ , Ar and Br are obtained directly from A and
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
2 3
 TK30 1 Kqdc 2 3
6 d0 Td0 0 Td0 0 7 0:3965 0:1983 0:1212
6 7 6 7
Ar ¼ 6  KA K6  1  KTvdc 7 ¼ 4 5483:2 100 2994:2 5;
4 TA TA A 5
K8 0 K9  C 1r 0:4067 0 100:0746
2 3 dc bess

K K 2 3
 T pw0  T pm0 0:0564 0:3594
6 7
6 7 6 7
d0 d0

Br ¼ 6  KA Kvw  KA Kvm 7 ¼ 4 138:495 8882 5;


4 TA TA 5
2:6858 0:2212
Kdw Kdm
2 3 2 3
K2 0:6702
6 7 6 7
Cr ¼ 4 0 5 ¼ 4 0 5; Dr ¼ ½ Kpw Kpm  ¼ ½ 0:7086 0:5355 
Kpdc 0:1805

It can have

DTet ¼ ½CTr ðsI  Ar Þ1 Br þ Dr u ¼ Fact ðsÞD/ þ Frct ðsÞDm ð4:132Þ


4.4 Examples 169

Hence,

½ Fact ðjxs Þ Frct ðjxs Þ  ¼ CTr ðjxs I  Ar Þ1 Br þ Dr


¼ ½ 0:7164\  0:136 0:5313\117:661 

The above results are as same as those given in Eq. (4.130). Take the transfer
function of the BESS stabilizer to be

ð1 þ sT2 Þ
Tbess ðsÞ ¼ Kbess with T1 ¼ 0:38s:
ð1 þ sT1 Þ

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 ¼ Fact ðjxs ÞD/ ¼ Fact ðjxs ÞTbess ðjxs ÞDPt


¼ Fact ðjxs ÞTbess ðjxs ÞðD þ jxs MÞDx ð4:133Þ

The electric torque supplied by a BESS reactive power stabilizer is

DTetreactive ¼ Frct ðjxs ÞDm ¼ Frct ðjxs ÞTbess ðjxs ÞDPt


¼ Frct ðjxs ÞTbess ðjxs ÞðD þ jxs MÞDx ð4:134Þ

The BESS active and reactive power stabilizer can be designed by using the phase
compensation method to supply a pure damping torque DTetactive ¼ 9Dx and
DTetreactive ¼ 20Dx, respectively. The parameters of stabilizers obtained from the
design are as follows:
1. The BESS active power stabilizers: Kbess ¼ 15:9567; T2bess ¼ 0:3826s:;
2. The BESS reactive power stabilizers: Kbess ¼ 99:7980; T2bess ¼ 0:0151s:
Table 4.1 gives the computational results of oscillation mode of example power
system without and with the BESS stabilizers installed. They confirm the effec-
tiveness 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 4 Single-Machine Infinite-Bus Power Systems …

Rotor angle δ (degree)


with reactive stabilizer installed

with active stabilizer installed


without stabilizers
time(second)

Fig. 4.20 Simulation results of example power system when Pt0 ¼ 0:5 p.u.

short circuit occurred at 0.5 s of simulation for 100 ms. The results confirm 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 ¼ 1:0; Vb ¼ 1:0; Vs0 ¼ 1:0 being fixed. 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 electrome-
chanical oscillation mode. Figures 4.22 and 4.23 are the results of simulation of
example power system at different load conditions. Those results obviously confirm
that the BESS active power stabilizer is more robust than the reactive power sta-
bilizer 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

with reactive stabilizer installed

with active stabilizer installed

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
Pt0 = 0.2 p.u. Pt0 = 0.9 p.u.
Without any stabilizer 0:0070  j6:3064 0:2505  j6:8670
With active power stabilizer 0:6685  j6:2676 0:5427  j6:8615
With reactive power stabilizer 0:2050  j5:7870 0:9353  j4:6027

Rotor angle δ (degree)


with active stabilizer installed
with reactive stabilizer installed

without stabilizers time(second)

Fig. 4.22 Simulation results of example power system when Pt0 ¼ 0:2 p.u.
172 4 Single-Machine Infinite-Bus Power Systems …

Rotor angle δ (degree)

with active stabilizer installed

with reactive stabilizer installed

without stabilizers
time(second)

Fig. 4.23 Simulation results of example power system when Pt0 ¼ 0:9 p.u.

4.4.2 An Example Single-Machine Infinite-Bus Power


System Installed with a UPFC Stabilizer

4.4.2.1 Extended Heffron–Phillips Model of Example Power


System Installed with a UPFC

Configuration of an example single-machine infinite-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 xd ¼ 1:0; xq ¼ 0:6; x0d ¼ 0:3; M ¼ 8s:; D ¼ 0; T0d0 ¼ 5:044s:
The AVR KA ¼ 100; TA ¼ 0:01s:
Transmission xtb ¼ 0:3; xte ¼ 0:15; xbt ¼ 0:15:
line
The UPFC ke ¼ 3; kb ¼ 3; Cdc ¼ 1:0; Vdc0 ¼ 1:0; xes ¼ 0:02; xbs ¼ 0:02:
Steady-state Pt0 ¼ 0:1; Vt0 ¼ 1:05; Vb ¼ 1:0:
operating point
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 ¼ 0; Vbt ¼ 0 which can give

Vbs ¼ jxbs Ibs ; Ves ¼ Vet ð4:135Þ

Total reactance of the transmission lines is xt ¼ xxtbtb ðþxxte teþþxxbtbtÞ ¼ 0:15. From


Vt0  Vb
Vb
jxt
¼ Pt0 þ jQb0 , the reactive power received at the infinite bus bar Qb0 can be
obtained as
2sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3

2
Vb 4 xt Pt0
Qb0 ¼ V2t0   Vb 5 ¼ 0:3326 ð4:136Þ
xt Vb

Thus,

Pt0  jQb0
Itb0 þ Ite0 ¼ ¼ 0:1  j0:3326
Vb
Vt0 ¼ jxt ðItb0 þ Ite0 Þ þ Vb ¼ 1:0499 þ j0:0150 ¼ 1:05\0:82
xtb
Ite0 ¼ Ibs0 ¼ ðItb0 þ Ite0 Þ ¼ 0:05  j0:1663 ¼ 0:1737\  73:27
xtb þ xte þ xbt
EQ ¼ Vt0 þ jxq ðItb0 þ Ite0 Þ ¼ 1:2495 þ j0:0750 ¼ 1:252\3:435
Ves0 ¼ Vet0 ¼ Vt0  jxte Ite0 ¼ 1:0249 þ j0:0075 ¼ 1:025\0:419 ¼ me0 ke Vdc0 \86:99
Vbs0 ¼ jxbs Ibs ¼ 0:0033  j0:001 ¼ 0:0034\16:86 ¼ mb0 kb Vdc0 \86:56

Hence, d0 ¼ 3:435 ; me0 ¼ 0:3417; mb0 ¼ 0:0012; de0 ¼ 0; db0 ¼ 2:8459 ,

vtd0 ¼ ReðVt0 Þ sin d0  ImðVt0 Þ cos d0 ¼ 0:0479;


vtq0 ¼ ReðVt0 Þ cos d0 þ ImðVt0 Þ sin d0 ¼ 1:0489;
ited0 ¼ ReðIte0 Þ sin d0  ImðIte0 Þ cos d0 ¼ 0:1690;
iteq0 ¼ ReðIte0 Þ cos d0 þ ImðIte0 Þ sin d0 ¼ 0:0399;
vtq0  Vb cos d0
itbd0 ¼ ¼ 0:1690;
xtb
Vb sin d0  vtd0
itbq0 ¼ ¼ 0:0399;
xtb
vetd0 ¼ ReðVet0 Þ sin d0  ImðVet0 Þ cos d0 ¼ 0:0539;
vetq0 ¼ ReðVet0 Þ cos d0 þ ImðVet0 Þ sin d0 ¼ 1:0236;
iesd0 ¼ 0; iesq0 ¼ 0;
ibsd0 ¼ ReðIbs0 Þ sin d0  ImðIbs0 Þ cos d0 ¼ 0:1690;
ibsq0 ¼ ReðIbs0 Þ cos d0 þ ImðIbs0 Þ sin d0 ¼ 0:0399;
E0q0 ¼ EQ  ðxd  x0d Þðited0 þ itbd0 Þ ¼ 1:1503
174 4 Single-Machine Infinite-Bus Power Systems …

Various coefficient matrices in Eqs. (4.86) and (4.88) are computed to be


   
0:9805 0 0:4001 0:0367 1:4073 7:9475
Fq ¼ Fd ¼ ;
0:0134 0 0:2689 0:0408 0:1656 5:0655
 
1:1742 7:6144 0:9001 0:0174
Gq ¼
0:7903 5:1251 0:9515 0:0184
 
23:3120 0:4196 14:8739 0:001
Gd ¼
14:8800 0:2678 15:8655 0:0011
Hd ¼ ½ 0:8589 0 0:0262 
Ld ¼ ½ 0:0768 0:4979 0:0103 0:0002 
Hq ¼ ½ 0:0416 0:2641 0:4323 
Lq ¼ ½ 1:2648 0:0228 0:1487 0

Following state equation of extended Heffron–Phillips model of Eq. (4.95) is


obtained
2 3 2 32 3
Dd_ 0 314:1593 0 0 0 Dd
6 7 6 76 7
6 Dx_ 7 6 0:2244 0 0:0173 0:0025 76 Dx 7
0
6 7 6 76 7
6 DE_ 7 6
0
6 q 7 ¼ 6 0:0192 0 0:5387 0:1983 0:2000 7 6 0 7
76 DEq 7
6 7
6 _0 7 6 76
2638:0 100 4330:5 54 DE0fd 5
7
4 DEfd 5 4 23:2301 0
DV_ dc 1:0061 0 0:0759 0 0:0189 DVdc
2 3
0 0 0 0 2 3
6 7 DmE
6 0:0074 0:1301 0:0012 0:00005 76
6 76 DdE 77
6 0:5851 0:0105 0:0688 0:000004 7 ð4:137Þ
6 764
7
6 7 Dm B5
4 12670 0:1463 1490:5 0:0127 5
DdB
0:0549 7:7987 0:0308 0:0172

Eigenvalues of state matrix are

k1 ¼ 94:4375
k2 ; k3 ¼ 0:0106  j8:4249
k4 ¼ 6:0714
k5 ¼ 0:0276

Obviously, k2 ; k3 ¼ 0:0106  j8:4249 are the pair of electromechanical


oscillation mode of example power system.
4.4 Examples 175

4.4.2.2 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 ¼ 1:05; Vb ¼ 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 ¼ 1; 2; 3; 4 where

Du1 ¼ Dme ; Du2 ¼ Dde ; Du3 ¼ Dmb ; Du4 ¼ Ddb ð4:138Þ



According to the definition of Eq. (4.108) and from Fig. 4.17, Fupfck ðjxs Þ ; k ¼
1; 2; 3; 4 is calculated as shown in Fig. 4.24 (jxs ¼ j8:4249).
From Fig. 4.24, it can be seen that

rselected ¼ fPt0 ¼ 0:1g ¼ Minð Fupfck ðjxs Þ ; k ¼ 1; 2; 3; 4; r 2 fPt0 ¼ 0:1  1:2gÞ
r

It means that the UPFC stabilizer should be designed at load condition Pt0 ¼ 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 ¼ Dde ¼ Max½ Fupfck ðjxs Þ ; k ¼ 1; 2; 3; 4; Duk 2 fDme ; Dmb ; Dde ; Ddb g
Duk

Magnitude of Forward path

with δ e signal

with m e signal

with m b signal
with δ b signal

Operating condition


Fig. 4.24 Computational result of Fupfck ðjxs Þ ; k ¼ 1; 2; 3; 4
176 4 Single-Machine Infinite-Bus Power Systems …

Hence, to have a robust UPFC stabilizer, the stabilizer should be added on Dde
and the design carried out at operating condition Pt0 ¼ 0:1; Vt0 ¼ 1:05; Vb ¼ 1:0.
The robust modulation signal to add a UPFC stabilizer can also be selected by
using the residue index defined by Eq. (4.104). Take the feedback signal of sta-
bilizer to be the locally available signal, the deviation of line active power Pte .
Similar to Eq. (4.83), it has

Pte ¼ vtq iteq þ vtd ited ð4:139Þ

By using Eqs. (4.86) and (4.88), similar to the derivation of Eq. (4.91), lin-
earization of Eq. (4.139) can be obtained to be
DPte ¼ iteq0 Dvtq þ ited0 Dvtd þ vtq0 ðDiesq þ Dibsq Þ þ vtd0 ðDiesd þ Dibsd Þ
2 3
Dd 2 3
6 7 Dme
6 Dx 7 6 Dd 7
6 7 6 e7 ð4:140Þ
¼ ½ c1 0 c3 0 c5  6 0 7
6 DEq 7 þ ½ d1 d2 d3 d4 6 7
6 7 4 Dmb 5
4 DE0fd 5
Ddb
DVdc

where

½ c1 0 c3 0 c5  ¼ Itq0 Hq þ Itd0 Hd þ Vtq0 Fq1 þ Vtq0 Fq2


þ Vtd0 Fd1 þ Vtd0 Fd2 ¼ ½ 1:1616 0 0:0859 0 0:0212 
½ d1 d2 d3 d4  ¼ Itq0 Lq þ Itd0 Ld þ Vtq0 Gq1 þ Vtq0 Gq2
þ Vtd0 Gd1 þ Vtd0 Gd2 ¼ ½ 0:0620 2:7016 0:0141 0:0011 

Fq1 and Fq2 , Fd1 and Fd2 , Gq1 and Gq2 and Gd1 and Gd2 are the first and
second rows of Fq , Fd , Gq , and Gd , respectively.
At the selected operating condition rselected ¼ fPt0 ¼ 0:1; Vt0 ¼ 1:05;
Vb0 ¼ 1:0g, for the oscillation mode k2;3 ¼ 0:0106  j8:4249, the right and left
eigenvectors are obtained as
2 3 2 3
0:0091 þ j0:0206 0:0091  j0:0206
6 7 6 7
6 0:7702 þ j0:3380 7 6 0:7702  j0:3380 7
6 7 6 7
v2 ¼ 6 7 6 7
6 0:0013 þ j0:0008 7 and v3 ¼ 6 0:0013  j0:0008 7
6 7 6 7
4 0:0000 þ j0:0000 5 4 0:0000  j0:0000 5
0:0010  j0:0010 0:0010 þ j0:0010
2 3 2 3
0:0899  j0:2015 0:0899 þ j0:2015
6 7 6 7
6 0:0054  j0:0024 7 6 0:0054 þ j0:0024 7
6 7 6 7
2 ¼ 6
w 7
6 0:0021  j0:0224 7 and w 3 ¼ 6
6 0:0021 þ j0:0224 7
7
6 7 6 7
4 0:9747 5 4 0:9747 5
0:0243 þ 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 ¼ 1; Du1 ¼ DmE ;
  T2 b1 ¼ 2:4108  j2:3637;
b21 ¼ w b31 ¼ w
 T3 b1 ¼ 2:4108 þ j2:3637
k ¼ 2; Du2 ¼ DdE ;

b22 ¼  T2 b2
w ¼ 10:7641  j5:1591; b32 ¼ w
 T3 b2 ¼ 10:7641 þ j5:1591
k ¼ 3; Du3 ¼ Dmb ;

b23 ¼  T2 b3
w ¼ 0:2548  j0:2928; b33 ¼ w
 T3 b3 ¼ 0:2548 þ j0:2928
k ¼ 4; Du4 ¼ Ddb ;
  T2 b4 ¼ 0:0023  j0:0001;
b24 ¼ w b34 ¼ w
 T3 b4 ¼ 0:0023 þ j0:0001

where
2 3 2 3 2 3 2 3
0 0 0 0
6 0:0074 7 6 0:1301 7 6 0:0012 7 6 0:0000 7
6 7 6 7 6 7 6 7
b1 ¼ 6 0:5851 7; b2 ¼ 6 0:0105 7; b3 ¼ 6 0:0688 7; b4 ¼ 6
6 7 6 7 6 7 7
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 ¼ cTk 
v2 ¼ 0:0002  j0:0029 and c3k ¼ cTk v3 ¼ 0:0002 þ j0:0029

where
cTk ¼ ½ 1:1616 0 0:1387 0 0:0202  ðk ¼ 1; 2; 3; 4Þ

The residue index for four input signals fDmE ; DdE ; DmB ; DdB g of the UPFC
stabilizer, thus, is obtained as

k ¼ 1; Du1 ¼ DmE ;
R21 ¼ 
b21c21 ¼ 0:0073  j0:0065; R31 ¼ b31c31 ¼ 0:0073 þ j0:0065;
k ¼ 2; Du2 ¼ DdE ;
R22 ¼ 
b22c22 ¼ 0:0128 þ j0:0322; R32 ¼ b32c32 ¼ 0:0128  j0:0322;
k ¼ 3; Du3 ¼ Dmb ;
R23 ¼ 
b23c23 ¼ 0:0009  j0:0007; R33 ¼ b33c33 ¼ 0:0009 þ j0:0007;
k ¼ 4; Du4 ¼ Ddb ;
24c24 ¼ 0:0000 þ j0:0000; R34 ¼ b34c34 ¼ 0:0000  j0:0000
R24 ¼ 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 4 Single-Machine Infinite-Bus Power Systems …

Residue

with m e signal

with δ e signal

with m b signal
with δ b signal

Pt0

Fig. 4.25 Computational result of jRki j; k ¼ 1; 2; 3; 4


R21 ¼ R31 ¼ 0:0097

R22 ¼ R32 ¼ 0:035

R23 ¼ R33 ¼ 0:0011

R24 ¼ R34 ¼ 0

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

uselected ¼ dE where Rik max ¼ Ri2 ¼ 0:035ði ¼ 2; 3Þ

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 ¼ 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 ¼ 0:1; Vt0 ¼ 1:05; Vb ¼ 1:0 of the example power system.
With Dde being selected to add the damping control signal of the UPFC stabi-
lizer, it can be obtained that
4.4 Examples 179

k ¼ 0:2244; d ¼ 0; C1 ¼ 1:1616; C2 ¼ 0; C3 ¼ ½ 0:0859 0 0:0212 ;


dm ¼ 2:7016
A23 ¼ ½ 0:0173 0 0:0025 
2 3 2 3 2 3
0:0192 0 0:5387 0:1983 0:2
6 7 6 7 6 7
A31 ¼ 4 23:2301 5; A32 ¼ 4 0 5; A33 ¼ 4 2638:0 100 4330:5 5
1:0061 0 0:0759 0 0:0189
B2 ¼ 0:1301
2 3
0:0105
6 7
B3 ¼ 4 0:1463 5
7:7986

Results of computation according to Eq. (4.61) are (jxs ¼ j8:4249)


jxs jxs
Kðjxs Þ ¼ k þ d þ AT23 ðjxs I  A33 Þ1 A31 þ A32 ¼ 0:2259 þ j0:0006
x0 x0

x0 x
C1 þ C2 þ CT3 ðjxs I  A33 Þ1
0
Ko ðjxs Þ ¼ A31 þ A32 ¼ 0:0663  j43:5996
jxs jxs
Kc ðjxs Þ ¼ AT23 ðjxs I  A33 Þ1 B3 þ B2 ¼ 0:1418 þ j0:0047
Kil ðjxs Þ ¼ CT3 ðjxs I  A33 Þ1 B3 þ dm ¼ 2:7599 þ j0:0155

Let the design of the UPFC stabilizer provide a pure damping torque
Dupfc Dx ¼ 1Dx. Take DPte as the feedback signal. Thus,

ð1 þ sT2 Þ
Dde ¼ Tupfc ðsÞDPte ¼ Kupfc DPte ð4:141Þ
ð1 þ sT1 Þ

where T1 ¼ 0:3 s.
From Eq. (4.64), the forward path of the stabilizer is calculated as

Fupfc ðjxs Þ ¼ Kc ðjxs ÞKo ðjxs Þ þ Dupfc Kil ðjxs Þ ¼ 2:9764  j6:1674

Thus, parameters of the UPFC stabilizer are obtained as

Kupfc ¼ 0:2689; T2 ¼ 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 confirm that the
UPFC stabilizer is effectively designed in damping the power oscillation and robust
to the variations of system load conditions.
180 4 Single-Machine Infinite-Bus Power Systems …

Table 4.3 Oscillation mode of example power system with variable load conditions
Pt0 ¼ 0:1 p.u. Pt0 ¼ 0:8 p.u. Pt0 ¼ 1:2 p.u.
Without stabilizer 0:0106  j8:4249 0:0133  j8:9982 0:1457  j9:4250
With stabilizer 0:6880  j7:0257 1:2283  j7:5835 1:2930  j8:1842

δ (degree)

without stabilizer

with UPFC stabilizer


time(second)

Fig. 4.26 Simulation results of example power system when Pt0 = 0.1 p.u.

δ (degree)
with UPFC stabilizer
without stabilizer

time(second)

Fig. 4.27 Simulation results of example power system when Pt0 = 0.8 p.u.
4.4 Examples 181

δ (degree)

with UPFC stabilizer


without stabilizer

time(second)

Fig. 4.28 Simulation results of example power system when Pt0 = 1.2 p.u.

In order to demonstrate that the robustness of the UPFC stabilizer is due to the
selection of the load condition at Pt0 ¼ 0:1; Vt0 ¼ 1:05; Vb ¼ 1:0 where the
stabilizer is designed, the stabilizer is designed at another load condition
Pt0 ¼ 0:8; Vt0 ¼ 1:05; Vb ¼ 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 ¼ 0:8; Vt0 ¼ 1:05; Vb ¼ 1:0, where the power oscillation is
effectively suppressed (see the second column of Table 4.4), its effectiveness is
not ensured at Pt0 ¼ 0:1; Vt0 ¼ 1:05; Vb ¼ 1:0 (see the first column of
Table 4.4).
At the selected operating condition Pt0 ¼ 0:1; Vt0 ¼ 1:05; Vb ¼ 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 ¼ 0:8; Vt0 ¼ 1:05; Vb ¼ 1:0
Pt0 ¼ 0:1 p.u. Pt0 ¼ 0:8 p.u. Pt0 ¼ 1:2 p.u.
Without stabilizer 0:0106  j8:4249 0:0133  j8:9982 0:0133  j8:9982
With stabilizer 0:4291  j6:8578 0:7517  j7:3203 0:7565  j7:5281
182 4 Single-Machine Infinite-Bus Power Systems …

Table 4.5 Results of designing the UPFC stabilizers added on other two modulation signals
Modulation signal added with Parameters of the UPFC Oscillation mode with the
a UPFC stabilizer stabilizer stabilizer installed
Dme Kp ¼ 1:6654; Tp ¼ 0:0926 0:7030  j6:9400
Dmb Kp ¼ 15:3864; Tp ¼ 0:1069 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 efficient 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):521–527
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 unified power flow controller. IEE Proc Part C. (1)
7. Wang HF (2000) A unified model for the analysis of FACTS devices in damping power system
oscillations part III: unified 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 configuration 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 þ jxLi ; i ¼ 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 x–y coordinate

Ig1 ¼ 1  g1 þ 1   3 Þ ¼ ðgL1 þ jbL1 þ jb13 ÞV 3


 g1  jb13 V
V ðVg1  V
rL1 þ jxL1 jx13
ð5:1Þ
Ig2 ¼ 1  g2 þ 1   3 Þ ¼ ðgL2 þ jbL2 þ jb23 ÞV 3
 g2  jb23 V
V ðVg2  V
rL2 þ jxL2 jx23

© Springer Science+Business Media New York 2016 183


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_5
184 5 A Multi-machine Power System Installed …

Vg1 V3 Vg 2 G2
G1 Ig1 x13 x 23 Ig 2

rL1 + jx L1 rL3 + jx L3 rL2 + jx L2

Fig. 5.1 A two-machine power system

1  3 þ 1 ðV 3  V  g1 Þ þ 1 ðV 3  V  g2 Þ
0¼ V
rL3 þ jxL3 jx13 jx23
¼ ðgL3 þ jbL3 þ jb13 þ jb23 ÞV  g1  jb23 V
 3  jb13 V  g2 ð5:2Þ

where jb13 ¼ jx1 ; jb23 ¼ jx1 ; gLi þ jbLi ¼ r þ1jx ; i ¼ 1; 2; 3:


13 23 Li Li
Matrix form of above equations is
2 3 2 32  g1 3
Ig1 gL1 þ jbL1 þ jb13 0 jb13 V
4 Ig2 5 ¼ 4 0 gL2 þ jbL2 þ jb23 jb23 56
4  g2 7
V 5
0 jb13 jb23 gL3 þ jbL3 þ jb13 þ jb23 V3
ð5:3Þ

From Eq. (5.2), it can have

3 ¼  jb13  g1  jb23  g2
V V V ð5:4Þ
gL3 þ jbL3 þ jb13 þ jb23 gL3 þ jbL3 þ jb13 þ jb23

Substituting above equation into Eq. (5.1), it can be obtained that


2 3
" # jb jb
gL1 þ jbL1 þ jb13 þ g 13þ jb13
jb13 jb23 " #
Ig1 
6 R gL3 þ jbR 7 Vg1
¼4
L3
5
Ig2 jb23 jb13 jb23 jb23  g2
V
gL3 þ jbR g L2 þ jb L2 þ jb 23 þ gL3 þ jbR
" #" # " #

yN11 yN12 V g1  g1
V
¼ ¼Y N

yN21 yN22 V g2  g2
V
ð5:5Þ
5.1 Mathematical Model of a Multi-machine … 185

Fig. 5.2 Phasor diagram in qi


x–y coordinate and di qi y
coordinate

E qi '
(x qi − x di ')Iqi

jx di ' Igi

V gi
δi
x

di

where bLR ¼ bL3 þ b13 þ b23 and Y  N is the network admittance matrix with only
nodes of generator terminal left. Equation (5.2) can be rearranged as
2 3 2 32 3
igx1 gN11 bN11 gN12 bN12 vgx1
6 igy1 7 6 bN11 gN12 7 6 7
6 7 6 gN11 bN12 76 vgy1 7
4 igx2 5 ¼ 4 gN21 bN21 gN22 bN22 5 4 vgx2 5
ð5:6Þ
igy2 bN21 gN21 bN22 gN22 vgy2

where subscript x and y denotes x and y component of a variable in common x–y


coordinate, respectively. According to the phasor diagram of Fig. 2.9, in d–q
coordinate of the ith generator, di -qi , it can have (i ¼ 1; 2)

V  0  jx0 Igi þ ðxqi  x0 Þiqi


 gi ¼ E ð5:7Þ
qi di di

 0 is
From Fig. 5.2, it can be seen that in the common x–y coordinate, phase of E qi
0 
di and that of ðxiq  xid Þiqi is di  90 . Hence, in x–y coordinate, Eq. (5.7) is
(i ¼ 1; 2)

 gi ¼ E0 ejdi  jx0 Igi þ ðxqi  x0 Þiqi ejðdi 90 Þ


V ð5:8Þ
qi di di
186 5 A Multi-machine Power System Installed …

In matrix form, Eq. (5.8) is


" # " #" #  "  #
 g1
V ejd1 0 E0q1 jx0d1 0 Ig1
¼ 
 g2
V 0 e jd2 0
Eq2 0 jxd2 Ig2
0
" 
# 
ðxq1  x0d1 Þiq1 ejðd1 90 Þ 0 iq1
þ 0 jðd2 90 Þ
ð5:9Þ
0 ðxq2  xd2 Þiq2 e iq2

Denote

 gi ¼ E0 ejdi þ ðxqi  x0 Þiqi ejðdi 90 Þ ;


E i ¼ 1; 2 ð5:10Þ
qi di

as an “internal voltage” of generator. Equation (5.9) becomes


     0  
 g1
E  g1
V jxd1 0 Ig1
 g2 ¼  g2 þ 0 Ig2 ð5:11Þ
E V jx0d2

 gi ; i ¼ 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 ¼ 1; 2.
Substituting Eq. (5.5) into Eq. (5.11), it can be obtained that
    0 1    
Ig1 1 jxd1 0  g1
E  g1
E

¼ YN þ 
¼Y  ð5:12Þ
Ig2 0 jx0d2  g2
E Eg2

where matrix Y is the network admittance matrix with only internal voltage nodes
of generators left.

E g1 Vg1 V3 Vg 2 Eg 2
x d1 ' Ig1 x13 x 23 Ig 2 xd2 '

rL1 + jx L1 rL3 + jx L3 rL2 + jx L2

Fig. 5.3 Introduction of internal voltage nodes of generators


5.1 Mathematical Model of a Multi-machine … 187

That is ði ¼ 1; 2Þ

X
2 X
2
Igi ¼ igxi þ jigyi ¼  gj ¼
yij E

yij ½E0qj ejdj þ ðxqj  x0dj Þiqj ejðdj 90 Þ  ð5:13Þ
j¼1 j¼1

where   According to Fig. 5.2, in di -qi coordinate


yij ¼ yij ejaij is the element of Y.
Eq. (5.13) is (i ¼ 1; 2)

 X
2

Igi ¼ idi þ jiqi ¼ ðixi þ jiyi Þejð90 di Þ
¼

yij ½E0qj ejdj þ ðxqj  x0dj Þiqj ejðdj 90 Þ ejð90 di Þ

j¼1
X
2

¼ yij ½E0qj ejð90
 þ dj di Þ
þ ðxqj  x0dj Þiqj ejðdj di Þ 
j¼1
X
2

¼ yij ½E0qj ejð90 þ dj di þ aij Þ
þ ðxqj  x0dj Þiqj ejðdj di þ aij Þ 
j¼1
X
2
¼ yij ½E0qj cos(90 þ dj  di þ aij Þ þ jE0qj sin(90 þ dj  di þ aij Þ
j¼1
þ ðxqj  x0dj Þiqj cos(dj  di þ aij Þ þ j(xqj  x0dj Þiqj sin(dj  di þ aij Þ
X2
¼ yij ½E0qj sin(dj  di þ aij Þ þ jE0qj cos(dj di þ aij Þ
j¼1
þ ðxqj  x0dj Þiqj cos(dj  di þ aij Þ þ j(xqj  x0dj Þiqj sin(dj  di þ aij Þ
ð5:14Þ

From above equation, it can have ði ¼ 1; 2Þ

X
2
idi ¼ yij ½E0qj sinðdj  di þ aij Þ þ ðxqj  x0dj Þiqj cosðdj  di þ aij Þ
j¼1
ð5:15Þ
X
2
iqi ¼ yij ½E0qj cos(dj  di þ aij Þ þ ðxqj  x0dj Þiqj sin(dj  di þ aij Þ
j¼1

5.1.1.2 Linearized Model When the Classical Model of Synchronous


Generators Is Used

For the simplicity of discussion, firstly 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 ¼ 1; 2)
188 5 A Multi-machine Power System Installed …

d_ i ¼ xo ðxi  1Þ
1
x_ i ¼ ½Pmi  Pi  Di ðxi  1Þ
Mi
ð5:16Þ
Pi ¼ vdi idi þ vqi iqi
vdi ¼ xqi iqi
vqi ¼ E0qi  x0di idi

Hence, linearized classical model of generator is (i ¼ 1; 2)

Dd_ i ¼ xo Dxi
1
Dx_ i ¼ ðDPi  Di Dxi Þ ð5:17Þ
Mi
DPi ¼ ðvdi0  iqi0 x0di ÞDidi þ ðvqi0 þ idi0 xqi ÞDiqi

Linearization of Eq. (5.15) is (E0qi is a constant, i; j ¼ 1; 2)

Didi ¼ hdi ðDdj  Ddi Þ


Diqi ¼ hqi ðDdj  Ddi Þ ð5:18Þ
i 6¼ j

where hdi and hdi i ¼ 1; 2 are constants. Hence, from Eqs. (5.17) and (5.18), it can
have (i; j ¼ 1; 2)

DPi ¼ ki ðDdj  Ddi Þ ð5:19Þ

where ki i = 1, 2 are constants. State equation of the two-machine power system is


2 3 2 0 0 xo 0
32 3 2 3
Dd_ 1 Dd1 Dd1
6 _ 7 6 0 0 0 xo 776 Dd2 7 6 Dd 7
6 Dd2 7 6 76 7 6 27
6 7¼6 k Mk1 MD1 6 7 ¼ A 46 7 ð5:20Þ
4 Dx_ 1 5 6 0 74 Dx 5
1
4 M1 1 1 5 1 4 Dx1 5
Dx_ 2 Mk2 k2 0 MD2 Dx2 Dx2
2 M2 2

Equation (5.20) is equivalent to the following equation


0 2 3 2 31
D1 k1 k1  
0 M Dd1
@s2 I þ sxo 4 M1 5 þ xo 4 M1 1 5A
¼0 ð5:21Þ
0 D2 Mk2 k2 Dd2
M2 2 M2
5.1 Mathematical Model of a Multi-machine … 189

Characteristic equation of the system is


 
 2 D 1 þ x k1 k1 
 k þ kxo M xo M 
 1
oM
1 1 ¼0 ð5:22Þ
 k D k 
 xo M2 k þ kxo M2 þ xo M2 
2
2 2 2

Obviously, k ¼ 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 ¼ Dd1  Dd2 , the state equation of system of Eq. (5.20) becomes
2 3 2 0 xo 0
3
2 3 2 3
Dd_ 12 Dd12 Dd12
6 7 6  k1 MD1 7
4 Dx_ 1 5 ¼ 6
4 M1 1
0 74 Dx 5 ¼ A 4 Dx 5
5 1 3 1 ð5:23Þ
Dx_ 2 k2 0 D2
M Dx2 Dx2
M2 2

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 rel-
ative angular position. For example, if Dd1 ¼ Dd2 , DPi ¼ 0 ði ¼ 1; 2Þ, 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 Heffron–Phillips Model

Heffron–Phillips model of the two-machine power system installed with the PSS
can be established based on the following 4th-order model of synchronous gener-
ators ði ¼ 1; 2Þ

d_ i ¼ xo ðxi  1Þ
1
x_ i ¼ ½Pmi  Pi  Di ðxi  1Þ
Mi
0 1 ð5:24Þ
E_ qi ¼ 0 ðEqi þ Efd0i þ E0fdi Þ
Tdoi
0 1 0 KAi
E_ fdi ¼  E þ ðVrefgi  Vgi þ upssi Þ
TAi fdi TAi
190 5 A Multi-machine Power System Installed …

where

vid ¼ xqi iqi


vqi ¼ E0qi  x0di idi
Pi ¼ vdi idi þ vqi iqi ð5:25Þ
Eqi ¼ E0qi þ ðxdi  x0di Þidi
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Vgi ¼ v2di þ v2qi

Linearization of Eq. (5.15) is (i ¼ 1; 2)

X
2
Didi ¼ hdij DE0qj þ hdj ðDdj  Ddi Þ
j¼1

X
2 ð5:26Þ
Diqi ¼ hqij DE0qj þ hqj ðDdj  Ddi Þ
j¼1

i 6¼ j

By using Eq. (5.26), linearization of Eq. (5.25) can be obtained as

DP1 ¼ k11 ðDd2  Dd1 Þ þ k211 DE0q1 þ k212 DE0q2


DP2 ¼ k12 ðDd1  Dd2 Þ þ k221 DE0q1 þ k222 DE0q2
DEq1 ¼ k41 ðDd2  Dd1 Þ þ k311 DE00q1 þ k312 DE0q2
ð5:27Þ
DEq2 ¼ k42 ðDd1  Dd2 Þ þ k321 DE0q1 þ k322 DE0q2
DVg1 ¼ k51 ðDd2  Dd1 Þ þ k611 DE00q1 þ k612 DE0q2
DVg2 ¼ k52 ðDd1  Dd2 Þ þ k621 DE00q1 þ k622 DE0q2

By substituting Eq. (5.27) into the linearization of Eq. (5.24), it can have for G1

Dd_ 1 ¼ xo Dx1
1
Dx_ 1 ¼ ½k11 ðDd2  Dd1 Þ  k211 DE0q1  k212 DE0q2  D1 Dx1 
M1
0 1
DE_ q1 ¼ 0 ½k41 ðDd2  Dd1 Þ  k311 DE0q1  k312 DE0q2 þ DE0fd1 
Tdo1
0 1 KA1
DE_ fd1 ¼  DE0fd1 þ ½k51 ðDd2  Dd1 Þ  k611 DE0q1  k612 DE0q2 þ Dupss1 
TA1 TA1
ð5:28Þ
5.1 Mathematical Model of a Multi-machine … 191

and for G2

Dd_ 2 ¼ xo Dx2
1
Dx_ 2 ¼ ½k12 ðDd1  Dd2 Þ  k221 DE0q1  k222 DE0q2  D2 Dx2 
M2
0 1
DE_ q2 ¼ 0 ½k42 ðDd1  Dd2 Þ  k321 DE0q1  k322 DE0q2 þ DE0fd2 
Tdo2
0 1 KA2
DE_ fd2 ¼  DE0 þ ½k52 ðDd1  Dd2 Þ  k621 DE0q1  k622 DE0q2 þ Dupss2 
TA2 fd2 TA2
ð5:29Þ

Equations (5.28) and (5.29) are the Heffron–Phillips model of the two-machine
power system installed with PSS as shown in Fig. 5.4.
State-space representation of the Heffron–Phillips model is
2 : 32 32 3 2 3
Dd 0 xo I 0 0 Dd 0
6 D x: 7 6
6 7 6 M1 K1 M1 D M1 K2 0 7 6 7 6 7
76 Dx0 7 þ 6 0 7Dupss
6 _0 7¼4 54 DEq 5 4 0 5
4 DE q 5 T1d0 K4 0 T1
d0 K3 T1
d0
_ 0
DEfd T 1
A K5 KA 0 T1
A K5 KA T1
A DE0fd T1
A KA

ð5:30Þ

where
    " #  0 
Dd1 Dx1 DE0q1 DEfd1
Dd ¼ ; Dx ¼ ; ¼DE0q 0 ; DE 0
¼ ;
Dd2 Dx2 DEq2 fd
DE0fd2
     
k11 k11 k211 k212 k311 k312
K1 ¼ ; K2 ¼ ; K3 ¼ ;
k12 k12 k221 k222 k321 k322
     
k41 k41 k51 k51 k611 k612
K4 ¼ ; K5 ¼ ; K6 ¼ ;
k42 k42 k52 k522 k621 k622
     0 
M1 0 D1 0 Tdo1 0
M¼ ;D¼ ; Td0 ¼ ;
0 M2 0 D2 0 T0do2
     
KA1 0 TA1 0 Dupss1
KA ¼ ; TA ¼ ; Dupss ¼
0 KA2 0 TA2 Dupss2

The model can be shown in Fig. 5.5 which is the matrix form of Fig. 5.4.
192 5 A Multi-machine Power System Installed …

Δδ2

k1−1 k1−1

+
ΔE q 2 ' - 1 Δω1 ω0 Δδ1
k 2−12 Δδ2
- D1 + sM1 s
-
ΔE q 2 ' Δδ2
k 2−11
k 3−12 k 4−1 k 4−1 k 5−1 k 5−1
ΔE q1 '
- -+ +
1 K A1 + Δu pss1
+
k 3−11 + sTd01 ' ΔE fd1 ' 1 + sTA1 -
-
k 6−12
Δδ1 K 6−11
ΔE q 2 '
k1− 2 k1− 2

+
ΔE q1 ' - 1 Δω2 ω0 Δδ2
k 2− 21
- D 2 + sM 2 s Δδ1
-
ΔE q1 ' Δδ1
k 2− 22
k 3− 21 k 4− 2 k 4− 2 k 5− 2 k 5− 2
ΔE q 2 '
- -+ + -
1 + KA2 + Δu pss2
k 3− 22 + sTd02 ' ΔE fd 2 ' 1 + sTA 2 -
-

k 6− 22 k 6− 21

ΔE q1 '

Fig. 5.4 Heffron–Phillips 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 Heffron–Phillips 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 syn-
chronous generator is given by Eqs. (5.24) and (5.25) for i ¼ 1; 2; . . .; N.
Equation (5.8) for i ¼ 1; 2; . . .; N can be written in the following matrix form
5.1 Mathematical Model of a Multi-machine … 193

K1

-
ω0 I
(sM + D) −1
s
-
K4 K5
K2

- -
+ +
(K 3 + sTd0 ) −1 −1
(I + sTA ) K A Δu pss

K6

Fig. 5.5 Heffron–Phillips model of two-machine power system in matrix form


 g ¼ ejd E0  jx0 Ig þ ðx  x0 Þejðd90 Þ iq
V ð5:31Þ
q d q d

where
 
g ¼ V
V  gN T ; Ig ¼ Ig1 Ig2 . . . IgN T ;
 g2 . . . V
 g1 V
 T
E0q ¼ E0q1 E0q2 . . . E0q1 ; iq ¼ ½ iq1 iq2 . . . iq1  ;
T

 
ejd ¼ diagðejdi Þ; x0d ¼ diagðx0di Þ; ðxq  x0d Þejðd90 Þ ¼ diag½ðxqi  x0di Þejðdi 90 Þ 
ð5:32Þ

In Eq. (5.32), diag(ci Þ denotes an Nth-order diagonal matrix with ci being the ith
diagonal element.
Let the network equation be

Ig ¼ Y g
 NV ð5:33Þ

where Y N is the network admittance matrix with only the nodes of generator
terminal left. From Eqs. (5.31) and (5.33), it can have
h i
Ig ¼ Y
 ejd E0 þ ðxq x0 Þejðd90 Þ iq ¼ Y
 E g ð5:34Þ
q d

where Y ¼ ðY 1 þ jx0 Þ1 . The above equation gives the generator current in x–y
N d
coordinate as (i ¼ 1; 2; . . .; N)
194 5 A Multi-machine Power System Installed …

X
N
Igi ¼ 
yij ½E0qj ejdj þ ðxqj  x0dj Þiqj ejðdj 90 Þ  ð5:35Þ
j¼1

where   Hence, Eq. (5.15) can be obtained


yij ¼ yij ejaij is the elements of matrix Y.
for i ¼ 1; 2; . . .; N. Their linearization can be expressed in the following matrix
form

DId ¼ Fdd Dd þ Gdd DE0q þ Hdd DIq


ð5:36Þ
DIq ¼ Fqq Dd þ Gqq DE0q þ Hqq DIq

where
 T
Dd ¼ ½ Dd1 Dd2 . . . DdN T ; DE0q ¼ DE0q1 DE0q2 ... DE0qN ;

DId ¼ ½ Did1 Did2 ... DidN T ; DIq ¼ ½ Diq1 Diq2 ... DiqN T
f ddij ¼ yij ½E0qj0 cosðdj0  di0 þ aij Þ þ ðxqj  x0dj Þiqj0 sinðdj0  di0 þ aij Þ; i 6¼ j,
f qqij ¼ yij ½E0qj0 sinðdj0  di0 þ aij Þ  ðxqj  x0dj Þiqj0 cosðdj0  di0 þ aij Þ; i 6¼ j,
X
N X
N
f ddii ¼  f ddij ; f qqii ¼  f qqij ;
j¼1;j6¼i j¼1;j6¼i

gddij ¼ yij sinðdj0  di0 þ aij Þ; gqqij ¼ yij cosðdj0  di0 þ aij Þ;
hddij ¼ yij ðxqj  x0dj Þ cosðdj0  di0 þ aij Þ; hqqij ¼ yij ðxqj  x0dj Þ sinðdj0  di0 þ aij Þ

and f ddij ; f qqij ; gddij ; gqqij ; hddij and hqqij are the elements of matrix Fdd ; Fqq ; Gdd ;
Gqq ; Hdd and Hqq , respectively.
From Eq. (5.45), it can be obtained that

DId ¼ Fd Dd þ Gd DE0q
ð5:37Þ
DIq ¼ Fq Dd þ Gq DE0q

where

Fd ¼ Fdd þ Hdd Fq ; Gd ¼ Gdd þ Hdd Gq ;


1
Fq ¼ ðI  Hqq Þ Fqq ; Gq ¼ ðI  Hqq Þ1 Gqq
5.1 Mathematical Model of a Multi-machine … 195

Linearization of Eqs. (5.24) and (5.25) is

Dd_ i ¼ xo Dxi
1
Dx_ i ¼ ðDPi  Di Dxi Þ
Mi
0 1 ð5:38Þ
DE_ qi ¼ 0 ðDEqi þ DE0fdi Þ
Tdoi
0 1 KAi
DE_ fdi ¼  DE0 þ ðDVgi þ Dupssi Þ
TAi fdi TAi

Dvdi ¼ xqi Diqi


Dvqi ¼ DE0qi  x0di Didi
DPi ¼ vdi0 Diid þ vqi0 Diqi þ idi0 Dvdi þ iqi0 Dvqi ð5:39Þ
DEqi ¼ DE0qi þ ðxdi  x0di ÞDidi
vgdi0 vqi0
DVgi ¼ Dvgdi þ Dvgqi
Vgi0 Vgi0

In matrix form, the above equations are

Dd_ ¼ xo IDx
Dx_ ¼ M1 ðDP  DDxÞ
0 ð5:40Þ
DE_ q ¼ T01 0
d0 ðDEq þ DEfd Þ
0
DE_ fd ¼ T1 0 1
A DEfd þ TA KA ðDVg þ Dupss Þ

DP ¼ Iq0 DE0q þ ðVd0  Iq0 X0d ÞDId þ ðVq0 þ Id0 Xq ÞDIq


DEq ¼ DE0q þ ðXd  X0d ÞDId ð5:41Þ
DVg ¼ V01 0
g0 Vgq0 DEq  V01 0
g0 Vgq0 Xd DId þ V01
g0 Vgd0 Xq DIq

where
 T
Dx ¼ ½ Dx1 Dx2 ... DxN T ; DE0fd ¼ DE0fd1 DE0fd2 . . . DE0fdN ;
DP ¼ ½ DP1 DP2 ... DPN T ; DEq ¼ ½ DEq1 DEq2 ... DEqN T ;
T T
DVg ¼ ½ DVg1 DVg2 . . . DVgN  ; Dupss ¼ ½ Dupss1 Dupss2 . . . DupssN  ;
M ¼ diag(Mi Þ; D ¼ diag(Di Þ; T0d0 ¼ diag(T0d0i Þ; TA ¼ diag(TAi Þ; KA ¼ diag(KAi Þ;
Id0 ¼ diag(idi0 Þ; Iq0 ¼ diag(iqi0 Þ; Vd0 ¼ diag(vdi0 Þ; Vq0 ¼ diag(vqi0 Þ; Vg0 ¼ diag(Vgi0 Þ;
X0d ¼ diag(x0di Þ; Xq ¼ diag(xqi Þ; Xd ¼ diag(xdi Þ
196 5 A Multi-machine Power System Installed …

Substituting Eq. (5.37) into Eq. (5.41), it can have

DP¼K1 Dd þ K2 DE0q
DEq ¼K3 DE0q þ K4 Dd ð5:42Þ
DVg ¼K5 Dd þ K6 DE0q

By substituting Eq. (5.42) into Eq. (5.40), Heffron–Phillips model of the


N-machine power system with PSS installed can be obtained as
2 3 2 32 3 2 3
Dd_ 0 xo I 0 0 Dd 0
6 Dx_ 7 6
6 7 6 M1 K1 M1 D M1 K2 0 7 6 7 6 7
76 Dx0 7 þ 6 0 7Dupss
6 DE_ 0 7¼4 54 DEq 5 4 0 5
4 q 5 T1
d0 K4 0 T1
d0 K3 T1
d0
DE_
0
T1
A K5 KA 0 T1
A K5 KA T1
A DE0fd T1
A KA
fd

ð5:43Þ

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 ¼ 1; 2; . . .; N)

Dw_ di ¼ Dvdi þ rai Didi þ x0 Dwqi þ wqi0 Dxi


Dw_ qi ¼ Dvqi þ rai Diqi  x0 Dwdi  wdi0 Dxi
Dw_ fi ¼ Dvfi  rfi Difi ð5:44Þ
Dw_ Di ¼ rDi DiDi
Dw_ Qi ¼ rQi DiQi
2 3 2 31 2 3
Didi xdi xadi xadi Dwdi
6 7 6 7 6 7
4 Difi 5 ¼ 4 xadi xfi xadi 5 4 Dwfi 5
DiDi xadi xadi xDi DwDi ð5:45Þ
" # " #1 " #
Diqi xqi xaqi Dwqi
¼
DiQi xaqi xQi DwQi
5.1 Mathematical Model of a Multi-machine … 197

Dd_ i ¼ x0 Dxi
1
Dx_ i ¼  ðDTei þ Di Dxi Þ ð5:46Þ
Mi
DTei ¼ wdi0 Diqi  wqi0 Didi þ iqi0 Dwdi  idi0 Dwqi

1 KAi
Dv_ 0fi ¼  Dv0  ðDVgi þ Dupssi Þ
TAi fi TAi
vgdi0 vgqi0 ð5:47Þ
DVgi ¼ Dvgdi þ Dvgqi
Vgi0 Vgi0

In matrix form, Eqs. (5.44) to (5.47) can be arranged as

X_ g ¼ Ag Xg þ Bg Dupss þ Bgv DVdq


ð5:48Þ
DIdq ¼ Cg Xg

where
 T  T
Dwd ¼ Dwd1 Dwd2 . . . DwdN ; Dwq ¼ Dwq1 Dwq2 . . . DwqN ;
Dwf ¼ ½ Dwf1 Dwf2 . . . DwfN T ; DwD ¼ ½ DwD1 DwD2 . . . DwDN T ;
 T
DwQ ¼ DwQ1 DwQ2 . . . DwQN ; Dv0f ¼ ½ Dv0f1 Dv0f2 . . . Dv0fN T ;
h iT
Xg ¼ DdT DxT DwTd DwTq DwTf DwTD DwTQ Dv0T f

Dupss ¼ ½ Dupss1 Dupss2 ... DupssN T


DVdq ¼ ½ Dvgd1 Dvgq1 Dvgd2 Dvgq2 ... ... DvgdN DvgqN T
DIdq ¼ ½ Did1 Diq1 Did2 Diq2 ... ... DidN DiqN T

Conversion equations between x–y and di qi coordinate systems as shown in


Fig. 5.2 are (i ¼ 1; 2; . . .; N)
         
f di sin di cos di f xi f sin di cos di f di
¼ ; xi ¼ ð5:49Þ
f qi cos di sin di f yi f yi cos di sin di f qi

where f i is a phasor (i ¼ 1; 2; . . .; N). Linearization of above equations is


      
Didi sin di0 cos di0 Dixi ixi0 cos di0 þ iyi0 sin di0
¼ þ Ddi
Diqi cos di0 sin di0 Diyi ixi0 sin di0 þ iyi0 cos di0
      
Dvgdi sin di0 cos di0 Dvgxi vgxi0 cos di0 þ vgyi0 sin di0
¼ þ Ddi
Dvgqi cos di0 sin di0 Dvgyi vgxi0 sin di0 þ vgyi0 cos di0
ð5:50Þ
198 5 A Multi-machine Power System Installed …

In matrix form, above equation can be written as

DIdq ¼ Tg0 DIxy þ BgI Xg


ð5:51Þ
DVdq ¼ Tg0 DVxy þ BgV Xg

where

DVxy ¼ ½ Dvgx1 Dvgy1 Dvgx2 Dvgy2 ... . . . DvgxN DvgyN T


DIxy ¼ ½ Dix1 Diy1 Dix2 Diy2 ... . . . DixN DiyN T

In the x–y coordinate, network equation of Eq. (5.33) can be written as


2 3 2      32 v 3
ix1 gN11 bN11 gN12 bN12 gN1N bN1N gx1
6 iy1 7 6 ... 6 vgy1 7
6 7 6 bN11 gN11 b gN12 b gN1N 7 6 7
6 7    N12   N1N 7 6 7
6 ix2 7 6 gN21 bN21 bN22 bN2N 767 6 v 7
6 7 66
gN22
...
gN2N
76 v 7
gx2
6 iy2 7 6 bN21 g gN2N 76 gy2 7
6 7¼ bN22 gN22 bN2N 7
6 . 7 6 76
N21
6 .. 7 6 . .. .. .. 76 .. 7
6 7 6 .. . . . 7
 76 . 7 7
6 7 6    
bNNN 54 vgxN 7
6
4 ixN 5 4 gNN1 bNN1 gNN2 bNN2 gNNN 5
...
iyN bNN1 gNN1 bNN2 gNN2 bNNN gNNN vgyN
ð5:52Þ

where   N.
yNij ¼ gNij þ jbNij is the elements of network admittance matrix Y
Linearization of Eq. (5.52) is

DIxy ¼ YG DVxy ð5:53Þ

where
2     3
gN11 bN11 gN12 bN12 gN1N bN1N
6 
6 b gN11 b gN12 b gN1N 7
6  N11   N12   N1N 7
6 gN21 bN21 gN22 bN22 gN2N bN2N 77
6  7
YG ¼ 6
6
bN21 gN21 bN22 gN22 bN2N gN2N 7
7
6 .. .. .. .. 7
6 . . 7
6    .  . 7
4 gNN1 bNN1 gNN2 bNN2 gNNN bNNN 5

bNN1 gNN1 bNN2 gNN2 bNNN gNNN

From Eqs. (5.48), (5.51), and (5.53), it can have

X_ g ¼ Ag Xg þ Bg Dupss þ Bgv ðTg0 DVxy þ BgV Xg Þ


ð5:54Þ
Cg Xg ¼ Tg0 YG DVxy þ BgI Xg
5.1 Mathematical Model of a Multi-machine … 199

Hence, state equation of the power system is

X_ g ¼ AXg þ Bg Dupss ð5:55Þ

where

A ¼ Ag þ Bgv ½Tg0 ðTg0 YG Þ1 ðCg  BgI Þ þ 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 definition on eigenvalues ki , corresponding left and
right eigenvectors, vi and wTi , of state matrix A is given

Avi ¼ ki vi ;
ð5:56Þ
wTi A ¼ wTi ki ; i ¼ 1; 2; . . .; M

Also following two matrices are constructed by eigenvectors as


2 3
wT1
6 wT 7
6 27
V ¼ ½ v1 v2 . . . vn ; V1 ¼6 7
6 .. 7 ¼ W
T
ð5:57Þ
4 . 5
wTn

Equations (2.61) and (2.66) establish the following relationship between the state
variables xi ðtÞ; i ¼ 1; 2; . . .; M and modes of the system zi ðtÞ ¼ zi ð0Þeki t ; i ¼
1; 2; . . .; M to be

xk ðt) ¼ vk1 z1 ð0Þek1 t þ vk2 z2 ð0Þek2 t þ    þ vkn zn ð0ÞekM t ð5:58Þ


200 5 A Multi-machine Power System Installed …

zi ðt) ¼ w1i x1 ðt) þ w2i x2 ðt) þ    þ wMi xM ðt) ð5:59Þ

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 ðt) ¼ zi ð0Þeki t ; i ¼ 1; 2; . . .; M, contributes to the kth state variable
xk ðt). 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 ðt) on the ith mode zi ðt) ¼ zi ð0Þeki t ; i ¼
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 defined to be the
participation factor

jvki wki j
pki ¼ PM ð5:60Þ
i¼1 jvki wki j

If it is assumed that xk ð0Þ ¼ 1; xj ð0Þ ¼ 0 (j 6¼ k, j ¼ 1; 2; . . .; n), from


Eq. (5.59), it can have

zi ð0Þ ¼ wki ; i ¼ 1; 2; . . .; M ð5:61Þ

By substituting Eq. (5.61) into Eq. (5.58) it can have

X
n
xk ðt) ¼ vki wki eki t ð5:62Þ
i¼1

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 defined by Eq. (5.60) is named as the
participation factor.
As it is pointed out previously in Chap. 2, the electromechanical power oscil-
lations are closely related to the rotor motion equation of generators where state
variables are the deviation of rotor position Ddi and speed Dxi ; i ¼ 1; 2; . . .; M in
the state Eqs. (5.43) or (5.55). Hence for a pair of conjugate complex eigenvalue, k i ,
of state matrix, if the associated participation factors of state variables Ddi and
Dxi ; i ¼ 1; 2; . . .; M are much higher than those of other state variables, this means
i is the dominant oscillation mode responsible for the electromechanical oscilla-
k
tion. Hence, ki is an electromechanical oscillation mode. That is, if
5.2 Modal Analysis and Control of Power System … 201

!
p ðxk ¼ DdÞ þ pki ðxk ¼ DxÞ X
ri ¼ Pki  1; pki ¼ 1 ð5:63Þ
k pki ðxk 6¼ Dd and xk 6¼ DxÞ k

i is an electromechanical oscillation mode. ri defined by Eq. (5.63) is called the


k
correlation ratio of electromechanical loop, which measure how much closely the
eigenvalue k i is related to the rotor motion of generators (hence the electrome-
chanical oscillation). ri is often used to identify whether an eigenvalue k i is an
oscillation mode or not.
For the oscillation mode of interests, k i , let its corresponding (right) eigenvector
be vi ¼ ½  v1i  v2i . . . vMi  . Assume the rotor speed of generators, Dxj
T

(j ¼ 1; 2; . . .; N), to be the (N + 1)th to (2N)th state variables of the system of


Eq. (4.1), i.e.,

Dxj ¼ Dxj þ N ; j ¼ 1; 2; . . .; N ð5:64Þ

Equation (5.58) can be written as


2 3 2 3
x1 ðt) v1i
6 .. 7 6 .. 7
6 . 7 6 . 7
6 7 6 7
6 Dx1 7 6 vN þ 1i 7
6 7 6 7
6 Dx2 7 6 vN þ 2i 7
6 7 6 7 
6 .. 7 ¼ v1 z1 ð0Þe þ    þ 6 .. 7zi ð0Þeki t þ   
k1 t
ð5:65Þ
6 . 7 6 . 7
6 7 6 7
6 DxN 7 6 v2Ni 7
6 7 6 7
6 .. 7 6 .. 7
4 . 5 4 . 5
xM ðt) vMi

The (N + 1)th to (2N)th elements of vi , vki ¼ vki \uki ; k = N + 1, N +


2; . . .; 2N describe the mode shape of oscillation mode ki , because vki ¼ vki \u
ki
weighs the contribution of the oscillation mode to the rotor motion of involved
generators. By sketching the phasor diagram of vki ¼ 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 ¼ n þ jxi ; vN þ 1i ¼ jvN þ 1i j\0 ; vN þ 2i ¼ jvN þ 2i j\180


k ð5:66Þ
i

Equations (5.65) and (5.66) indicate that the rotor motion of the 1st and 2nd
generator, Dx1 and Dx2 , must have component jvN þ 1i jzi ð0Þeni cos xi t and
jvN þ 2i jzi ð0Þeni cosðxi t þ 180 Þ, respectively. Hence, as far as oscillation mode
i ¼ n þ jxi is concerned, the 1st and 2nd generators oscillate against each
k i
other. Thus, by looking at vki ¼ vki \uki ; k ¼ N þ 1; N þ 2;. . .; 2N, how groups
202 5 A Multi-machine Power System Installed …

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. Identification of electromechanical oscillation modes by checking the imaginary
part of eigenvalues within the range 2pf o ðf o ¼ 0:1  2 HzÞ 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 electrome-
chanical oscillation modes are needed. In addition, for a large-scale power system
with a high-dimensional state matrix, numerical difficulty of QR method may occur.
Hence, a lot of effort has been spent by many researchers to propose and develop
numerical methods for the efficient 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 difficulty. 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 Heffron–Phillips model.
Heffron–Phillips model of N-machine power system of Eq. (5.43) can be written
as (without PSS being considered)

sDd ¼ xo Dx
sDx ¼ M1 ðK1 Dd  K2 DE0q  DDxÞ
ð5:67Þ
sDE0q ¼ T01 0 0
d0 ðK3 DEq  K4 Dd þ DEfd Þ

sDE0fd ¼ T1 0 1 0
A DEfd þ TA KA ðK5 Dd  K6 DEq Þ
5.2 Modal Analysis and Control of Power System … 203

From Eq. (5.67), it can be obtained that

fs2 M þ sD þ xo K1  K2 ½ðI þ sTA ÞðK3 þ sT0d0 Þ þ KA K6 1


ð5:68Þ
½ðI þ sTA ÞK4 þ KA K5 gDd ¼ FðsÞDd ¼ 0

 is an electromechanical oscillation mode, it must have


Obviously, if ki
i Þ ¼ 0
DET½Fðk ð5:69Þ

where DET½M ¼ 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 elec-
tromechanical 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)
    
Dd_ 0 xo I Dd
¼ ð5:70Þ
Dx_ K1 0 Dx

The characteristic equation of the system is


  
 
kI  0 xo I  ¼ jkI  Ar j ¼ 0 ð5:71Þ
 K1 0 

The equivalent form of Eq. (5.71) is

ðs2 M þ xo K1 ÞDd ¼ ðs2 I þ xo M1 K1 ÞDd ¼ 0 ð5:72Þ

Hence, 2N eigenvalues of system state matrix Ar are also solutions of the


following equation
2 
s I þ xo M1 K1  ¼ 0 ð5:73Þ

Therefore, if N eigenvalues of matrix xo M1 K1 are gj ; j ¼ 1; 2;. . .; N, 2N


pffiffiffiffiffi
eigenvalues of matrix Ar are  gj ; j ¼ 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
pffiffiffiffiffi
j gj ; j ¼ 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.
pffiffiffiffiffi
gj 6¼ 0; j ¼ 1; 2; . . .; N  1; are the N − 1 angular natural oscillation frequency
of electromechanical oscillation modes.
For a poorly damped oscillation mode of interests, k i ¼ n  jxi , because it is
i
close to the imaginary axis of the complex plan and the angular oscillation
204 5 A Multi-machine Power System Installed …

Fig. 5.6 Optimal direct


one-dimensional searching y
along the direction of
imaginary and real axis on
Contour of f (λ )
complex plane
λi

0 + j ηi

pffiffiffiffiffi 
frequency xi is close to the angular natural oscillation frequency, gi , k i ð0Þ ¼
pffiffiffiffiffi i ¼ n  jxi .
0 þ j gi should be a good initial guess of oscillation mode, k i
Therefore, following objective function can be established

fðkÞ ¼ jDET½FðkÞj ð5:74Þ

A direct searching method can be used to find the minimal optimum of above
pffiffiffiffiffi
objective function on the complex plan, starting from 0 þ j gi . That will be the
i ¼ n  jxi , which satisfies 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 ¼ n  jxi is very close to 0 þ jpffiffiffiffi
as shown in Fig. 5.6. If k
ffi 
gi , k i ¼ 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 first
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
5.2 Modal Analysis and Control of Power System … 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 elec-
tromechanical 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
interests ki ¼ n  jxi is selected to be damped by the PSS installed on the jth
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
be used to assign the ith oscillation mode k  ¼
i ¼ n  jxi to a target position k
i i
n i  jx i by setting the parameters of the jth PSS. Afterwards, the (i + 1)th oscil-
lation 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 simulta-
neous 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,

fðpÞ; p 2 P ð5:75Þ

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 find 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 classified into two categories: (1) mathematical methods and (2) heuristic
methods. Mathematical methods include linear programming, nonlinear program-
ming 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 5 A Multi-machine Power System Installed …

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 find 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
Heffron–Phillips 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


 ¼ n  jx , i ¼ 1; 2; . . .; L be
i ¼ n  jxi , to target positions k
oscillation modes, k i i i i

X
L

fðpÞ ¼ i ðpÞk
½k  ðp Þ2 ð5:76Þ
i
i¼1

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

pðk þ 1Þ ¼ pðkÞ  st  rf½pðkÞ ð5:77Þ

where st is the optimal one-dimensional searching length and rf½pðkÞ 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.
Heffron–Phillips model of N-machine power system of Eq. (5.43) can be written
as (with PSSs being considered)

sDd ¼ xo IDx
sDx ¼ M1 ðK1 Dd  K2 DE0q  DDxÞ
ð5:78Þ
sDE0q ¼ T01 0 0
d0 ðK3 DEq  K4 Dd þ DEfd Þ

sDE0fd ¼ T1 0 1 0
A DEfd þ TA KA ðK5 Dd  K6 DEq þ Dupss Þ

From Eq. (5.78), it can be obtained that

fs2 M þ sD þ xo K1  K2 ½ðI þ sTA ÞðK3 þ sT0d0 Þ þ KA K6 1


½ðI þ sTA ÞK4 þ KA K5 gDd þ K2 ½ðI þ sTA ÞðK3 þ sT0d0 Þ ð5:79Þ
1
þ KA K6  KA Dupss ¼ 0
5.2 Modal Analysis and Control of Power System … 207

Let the transfer function of the jth PSS be Gj ðsÞ and the feedback signal Dxj .
Dupss can be denoted as
s
Dupss ¼ Gpss ðp; sÞDd ð5:80Þ
x0

where Gpss ðp; sÞ ¼ diagðGpssj Þ, Gpssj ¼ Gj ðs) when the jth generator is installed
with a PSS; otherwise, Gpssj ¼ 0. From Eqs. (5.79) and (5.80), it can have

fs2 M þ sD þ xo K1  K2 ½ðI þ sTA ÞðK3 þ sT0d0 Þ þ KA K6 1


s ð5:81Þ
½ðI þ sTA ÞK4 þ KA K5  KA Gpss ðp; sÞDdgDd ¼ Fðp; sÞDd ¼ 0
x0

Hence, it must have

DET½Fðp; ki Þ ¼ 0 ð5:82Þ

Equation (5.82) indicates that there should exist a nonzero vector vi and w
i
satisfying

Fðp; ki Þvi ¼ 0
ð5:83Þ
wTi Fðp; ki Þ ¼ 0

Let p ¼ ½ p1 p2 . . . ps T . A small increment of pk , Dpk , will cause that of ki


and 
vi , Dki and Dvi . It can have

F½pðpk þ Dpk Þ; ki þ Dki ðvi þ Dvi Þ ¼ 0 ð5:84Þ

Because

@Fðp; ki Þ @Fðp; ki Þ
F½pðpk þ Dpk Þ; ki þ Dki  ¼ Fðp; ki Þ þ Dki þ Dpk ð5:85Þ
@ki @pk

From Eqs. (5.83), (5.84), and (5.85), it can be obtained that

@ki  Ti @Fðp;k
w iÞ
@ pk vi

¼ ð5:86Þ
@pk T
w @Fðp;k i Þ
vi
i @ki

From Eq. (5.76), it can be obtained that

@ki
rf½pðkÞ ¼ 2ðki  k i Þ ð5:87Þ
@pðkÞ
208 5 A Multi-machine Power System Installed …

h iT
@ki @ki @ki @ki
where @pðkÞ
¼ @ p1 @ p2  @ ps and @@k
pk (k ¼ 1; 2; . . .; s) is calculated by
i

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 ðk þ 1Þ is very
close to ki ðkÞ and it should have DETF½pðk þ 1Þ; ki ðk þ 1Þ ¼ 0. Hence, set up the
following objective function,

f k þ 1 ðki Þ ¼ jDETF½p; ki j ð5:88Þ

The direct searching method introduced based on Eq. (5.74) can be used to find
ki ðk þ 1Þ on the complex plane, starting from ki ðkÞ.
In the multi-machine power system, objective of installation of multiple PSSs is
to ensure sufficient 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

fðpÞ ¼ minðfi Þ
0:05 ð5:89Þ

where

ki ¼ ni  jxi 2 fall oscillation modes to be damped by the PSS designg;


fi
fi ¼  qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ni þ x2i
2

is for the coordinated design of multiple PSSs to ensure the damping of all oscil-
lation 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 mul-
tiple PSSs can also be jointly used to achieve more efficient 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 ¼ 1; 2; . . .; L), to the electromechanical
oscillation loop of the jth generator. With this assumption, Eq. (5.81) becomes

fs2 M þ sðD þ Dpss Þ þ xo K1  K2 ½ðI þ sTA ÞðK3 þ sT0d0 Þ þ KA K6 1


ð5:90Þ
½ðI þ sTA ÞK4 þ KA K5 gDd ¼ Fðp; sÞDd ¼ 0

where
5.2 Modal Analysis and Control of Power System … 209

Dpss ¼ diag(dpssj Þ; dpssj ¼ Dpssj

when the jth generator is installed with a PSS; otherwise, dpssj ¼ 0 ðj ¼ 1; 2; . . .; L),
 T
p ¼ Dpssj Dpssðj þ 1Þ . . . Dpssðj þ LÞ .
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 ¼ 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 sys-
tem, 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 fixed 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
Heffron–Phillips model of N-machine power system installed with multiple PSSs.
Let the realization of the jth PSS, Dupssj ¼ Gj ðsÞDxj , be

Dupssj ¼ Hj zj ðtÞ þ kj Dxj


ð5:91Þ
z_ j ðtÞ ¼ Fj zj ðtÞ þ Sj Dxj

Hence, it can have

Dupss ¼ HzðtÞ þ Ky
ð5:92Þ
z_ ðtÞ ¼ FzðtÞ þ Sy

where
210 5 A Multi-machine Power System Installed …

 T T
zðtÞ ¼ z1 ðtÞT z2 ðtÞT    zL ðtÞT ; y ¼ ½ Dxj Dxj þ 1    Dxj þ L  ;
H ¼ blockdiagðHj Þ; F ¼ blockdiagðFj Þ; S ¼ blockdiagðSj Þ; K = diag(kj Þ

and blockdiagðMj Þ denotes a block diagonal matrix with the diagonal matrix to be Mj .
From Eq. (5.92), state equation of Heffron–Phillips model of N-machine power
system with L PSSs installed of Eq. (5.43) can be written as

_ ¼ AxðtÞ þ BDupss
xðtÞ
ð5:93Þ
y ¼ CxðtÞ

where
2 3
0 xo I 0 0
6 7
6 M1 K1 M1 D M1 K2 0 7
A¼6
6 T1 K
7;
7
4 d0 4 0 T1
d0 K3 T1
d0 5
T1
A K5 KA 0 T1
A K5 KA T1
A
2 3
0
6 7
6 0 7
6
B¼6 7; C ¼ ½ 0 I 0 0
7
4 0 5
T1
A KA

The set of fixed modes of the decentralized control system of Eqs. (5.92) and
(5.93), denoted as KðA; B; C; KÞ, is defined as

KðA; B; C; KÞ ¼ \ 0 k½A þ BKC ð5:94Þ


K2K

where k½A þ BKC denotes the set of all eigenvalues of matrix A þ BKC, and K0 is
the following set

K0 ¼ fKjK ¼ diagðkj Þ; K 2 RNN g ð5:95Þ

Definition of the fixed modes given by Eqs. (5.94) and (5.95) in fact means that
the fixed modes are not affected by variations of any kj . This is the key in the
following proof.
From Eq. (5.93), it can have
5.2 Modal Analysis and Control of Power System … 211

2 3
0 xo I 0 0
6 M1 K1 M1 D M1 K2 0 7
A þ BKC ¼ 6
4 T1 K4
7
5 ð5:96Þ
d0 0 T1
d0 K3 T1
d0
T1
A K5 KA
1 1
TA KA K TA K5 KA TA 1

If ko is a fixed mode, it must be the eigenvalue of matrix A þ BKC. Hence,


2 32 3 2 3
0 xo I 0 0 v1 v1
6 M K1 1
M 1
D M 1
K 0 7 6 7 6 7
6 2 76 2 7 ¼ ko 6 v2 7 ð5:97Þ
v
4 T1 K4 0 1
Td0 K3 1 54
Td0 v3 5 4 v3 5
d0
1 1 1 1
TA K5 KA TA KA K TA K5 KA TA v4 v4

That is,

xo v2 ¼ ko v1
 K1 v1  Dv2  K2 v3 ¼ ko Mv2
ð5:98Þ
 K4 v1  K3 v3 þ v4 ¼ ko Td0 v3
 K5 KA v1  KA Kv2  K6 KA v3  v4 ¼ ko TA v4

From Eq. (5.98), it can be obtained that


   
KA K
A4 k4o þ A3 k3o þ A2 k2o þ A1  k o þ A 0 v1 ¼ 0 ð5:99Þ
xo

where

A0 ¼ ðK3 þ KA K6 ÞK1
2 K1  K4  KA K5
1
A1 ¼ ðTA K3 þ Td0 ÞK1
2 K1 þ ðK þ KA K6 ÞK1
2 D  TA K4
x0 3
1 1
A2 ¼ ðTA K3 þ Td0 ÞK1 1
2 D þ TA Td0 K2 K1 þ ðK þ KA K6 ÞK1
2 M
x0 x0 3
1 1
A3 ¼ TA Td0 K1
2 Dþ ðT K3 þ Td0 ÞK1
2 M
x0 x0 A
1
A4 ¼ TA Td0 K1
2 M
x0

v 6¼ 0; hence, v1 6¼ 0. It should have


   
KA K
DET A4 k4o þ A3 k3o þ A2 k2o þ A1  ko þ A0 ¼ 0 ð5:100Þ
xo

Denote
212 5 A Multi-machine Power System Installed …

 
KA K
Fðko ; k1 ; k2; . . .kN Þ ¼ A4 k4o þ A3 k3o þ A2 k2o þ A1  ko þ A0
xo
2 3
f 11 ðko ; k1 Þ f 12 ðko Þ ... f 1N ðko Þ
6 f 21 ðko Þ f 22 ðko ; k2 Þ . . . f 2N ðko Þ 7 ð5:101Þ
6 7
¼6
6 .. .. .. ..
7
7
4 . . . . 5
f N1 ðko Þ f N2 ðko Þ ... f NN ðko ; kN Þ

Also denote
2 3
f 11 ðko ; k1 Þ f 12 ðko Þ ... f 1N ðko Þ
6 f 21 ðko Þ f 22 ðko ; k2 Þ . . . f 2N ðko Þ 7
6 7
Fk ðko ; k1 ; k2 ; . . .kN Þ ¼ 6 .. .. .. .. 7 ð5:102Þ
4 . . . . 5
f N1 ðko Þ f N2 ðko Þ ... f NN ðko ; kN Þ

Obviously, FN ðko ; k1 ; k2 ; . . .kN Þ ¼ Fðko ; k1 ; k2 ; . . .kN Þ. Let the companion


matrix of Fk ðko ; k1 ; k2; . . .kk Þ be F k ðko ; k1 ; k2; . . .kk Þ and denote
2 3
f k;11 f k;12 ... f k;1k
6 f k;21 f k;22 ...
f k;2k 7
6 7
Fk ðko ; k1 ; k2; . . .kk Þ ¼ 6 . .. .. .. 7 ð5:103Þ
4 .. . . . 5
f k;k1 f k;k2 ... f k;kk

From Eq. (5.102), it can have

f k;kk ¼ DET½Fk1 ðko ; k1 ; k2; . . .kk1 Þ ð5:104Þ

Denote

gk ðko ; k1 ; k2;    kk Þ ¼ DET½Fk ðko ; k1 ; k2;    kk Þ ð5:105Þ

When k ¼ N, from Eqs. (5.100) and (5.105) it can have

gN ðko ; k1 ; k2 ; . . .kN Þ ¼ DET½FN ðko ; k1 ; k2 ; . . .kN Þ


ð5:106Þ
¼ DET½Fðko ; k1 ; k2 ; . . .kN Þ ¼ 0

Because ko is a fixed mode,

@ko
¼ 0ðj ¼ 1; 2; . . .; NÞ ð5:107Þ
@kj

From Eq. (5.106), it can have


5.2 Modal Analysis and Control of Power System … 213

@gN @ko @g
þ N¼0 ð5:108Þ
@ko @kN @kN

From Eqs. (5.107) and (5.108), it can be obtained that

@gN
¼0 ð5:109Þ
@kN

For any matrix MðxÞ,

dDET½MðxÞ dMðxÞT
¼ Trace½M ðxÞ  ð5:110Þ
dx dx

where M ðxÞ is the companion matrix of MðxÞ; thus,

@gN @FN ðko ; k1 ; k2 ; . . .kN ÞT


¼ Trace½F N ðko ; k1 ; k2 ; . . .kN Þ ¼0 ð5:111Þ
@kN @kN

From Eq. (5.101), it can be obtained that

@FN ðko ; k1 ; k2 ;    kN ÞT h i
@ f NN ðko ;kN Þ
¼ diag 0 0  0
@kN @ kN
ð5:112Þ
h i
ko
¼ diag 0 0  0 x0 KNA

Hence from Eqs. (5.103), (5.111), and (5.112), it can have

@gN ko
¼ KNA f N;NN ¼ 0 ð5:113Þ
@kN x0

KNA 6¼ 0 and if ko is an electromechanical oscillation mode, ko 6¼ 0. Hence from


Eqs. (5.113), (5.104), and (5.105), it can have

f N;NN ¼ DET½FN1 ðko ; k1 ; k2 ; . . .kN1 Þ ¼ gN1 ðko ; k1 ; k2 ; . . .kN1 Þ ¼ 0 ð5:114Þ

From Eq. (5.114), it should have

@gN1 @ko @g
þ N1 ¼ 0 ð5:115Þ
@ko @kN1 @kN1

From Eqs. (5.101), (5.107), and (5.110), it can be obtained


214 5 A Multi-machine Power System Installed …

" #
@gN1 @FN ðko ; k1 ; k2 ; . . .kN1 ÞT
¼ Trace FN1 ðko ; k1 ; k2 ; . . .kN1 Þ
@kN1 @kN1
n h io
¼ Trace F N1 ðko ; k1 ; k2 ; . . .kN1 Þdiag 0 0 . . . kko0 KðN1ÞA
@o @o
¼ KðN1ÞA f ðN1Þ;ðN1ÞðN1Þ ¼ KðN1ÞA gN2 ð@o ; k1 ; k2 ; . . .kN2 Þ ¼ 0
@0 @0
ð5:116Þ

Again, KðN1ÞA 6¼ 0 and ko 6¼ 0, it can have

gN2 ðko ; k1 ; k2 ; . . .kN2 Þ ¼ DET½FN2 ðko ; k1 ; k2 ; . . .kN2 Þ ¼ 0 ð5:117Þ

Above procedure from Eqs. (5.106) to (5.117) can be carried on iteratively until
it is obtained that

@g1 ko
¼ K1A ¼ 0 ð5:118Þ
@k1 x0

It contradicts to that K1A 6¼ 0 and ko 6¼ 0. Hence, ko cannot be a fixed mode and


an electromechanical oscillation mode.

5.3 An Example Three-Machine Power System

5.3.1 Example Power System and Its Linearized


Heffron–Phillips Model

5.3.1.1 System Parameters and Operating Conditions

Configuration 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 ¼ 1:04\0; VG2 ¼ 1:025\0; VG3 ¼ 1:025\0;


V1;2;3 ¼ 1:00\0; VA;B;C ¼ 1:00\0;
SG1 ¼ 0:716 þ j0:27; SG2 ¼ 1:63 þ j0:067; SG3 ¼ 0:85  j0:109
LoadA ¼ 1:25 þ j0:5; LoadB ¼ 0:9 þ j0:3; LoadC ¼ 1 þ j0:35:

Tables 5.3 and 5.4 are the results of the load flow computation.
5.3 An Example Three-Machine Power System 215

0.716+j0.27 BusC 1.63 + j0.067


Bus1 1.25 + j0.5 Bus3

G1 G3
1.04∠0
1.025∠0

BusB BusC

0.9 + j0.3 1 + j0.35

Bus2

0.85 − j0.109
G2
1.025∠0

Fig. 5.7 Configuration of a three-machine example power system [1]

Table 5.1 Parameters of G1 G2 G3


generators and the AVRs [1,
2] xd 0.146 0.8958 1.313
x0d 0.0608 0.1189 0.1813
T0d0 8.96 6 5.89
xq 0.0969 0.8645 1.258
x0q 0.0969 0.1969 0.25
M 47.2 12.8 6.02
D 0 0 0
KA 200 200 200
TA 0.02 0.02 0.02

5.3.1.2 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 calcu-
its terminal voltage, V
lated as
216 5 A Multi-machine Power System Installed …

Table 5.2 Parameters of transformers and lines [1]


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

Table 5.3 Results of load flow computation—node voltage


Node Magnitude of voltage Phase of voltage (deg.)
G1 1.04 0.000
G2 1.025 0.163
G3 1.025 0.081
1 1.018 −0.039
A 0.977 −0.070
B 1.007 −0.065
C 1.012 0.012
2 1.021 0.065
3 1.030 0.034

Table 5.4 Results of load flow computation—active and reactive power


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

Igi ¼ Pi  jQi ; i ¼ 1; 2; 3

V gi

 Qi ; i ¼ 1; 2; 3, can be computed as
The imaginary voltage of each generator, E

 Qi ¼ Vgi þ jxqiIgi ;
E i ¼ 1; 2; 3

Phase of E Qi ; i ¼ 1; 2; 3 is the angular position of each generator, di , i ¼ 1; 2; 3.


From di , i ¼ 1; 2; 3, computational results of output current and terminal voltage of
each generator can be transformed from the common x–y coordinate to the d–q
coordinate as (Eq. (5.49))
        
idi sindi cosdi vgxi vdi sindi cosdi vgxi
¼ ¼
iqi cosdi sindi vgyi vqi cosdi sindi vgyi

According to Eq. (5.25), it can have

E0qi ¼ vqi þ x0di idi


Eqi ¼ E0qi þ (xdi  x0di Þidi

Since E0fdi0 ¼ 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 Heffron–Phillips Model

From Table 5.2 and Fig. 5.7, following network admittance matrix can be
established.
 
  12
 L ¼ Y11
Y
Y
 21
Y  22
Y

Table 5.5 Initial value of G1 G2 G3


state variables
di0 (rad.) 0.0621 1.0388 0.9199
xi0 1 1 1
E0qi0 1.0637 0.8119 0.7882
E0fqi0 1.0998 1.8302 1.4334
218 5 A Multi-machine Power System Installed …

where
2 3
j17:3611 0 0
 11
Y ¼4 0 j16 0 5;
0 0 j17:0648
2 3
j17:3611 0 0 0 0 0
 12
Y ¼ 4 0 0 0 0 j16 0 5;
0 0 0 0 0 j17:0648
2 3
j17:3611 0 0
6 0 0 0 7
6 7
6 7
 21 6 0 0 0 7
Y ¼6 7;
6 0 0 0 7
6 7
4 0 j16 0 5
0 0 j17:0648
2 3
3:3074  j39:3089 1:3652 þ j11:6041 1:9422 þ j10:5107
6 1:3652 þ j11:6041 3:8620  j18:0714 0 7
6 7
6 1:9422 þ j10:5107 4:1115  j16:1367 7
 221 6 0 7
Y ¼6 7;
6 0 0 0 7
6 7
4 0 1:1876 þ j 5:9751 0 5
0 0 1:2820 þ j5:5882
2 3
0 0 0
6 0 1:1876 þ j5:9751 0 7
6 7
6 1:2820 þ j5:5882 7
 222 6 0 0 7
Y ¼6 7;
6 3:7482  j23:6449 1:6171 þ j13:6980 1:1551 þ j9:7843 7
6 7
4 1:6171 þ j13:6980 2:8047  j35:4456 0 5
1:1551 þ j9:7843 0 2:4371  j32:1539

 22 ¼ ½Y
Y  222 
 221 Y

By keeping three nodes of generator terminals via deleting all the other nodes in
the above network admittance matrix, it can have

N ¼ Y
Y  11  Y  1 Y
 12 Y 
22 21
2 3
1:0994  j4:7578 0:0954 þ j2:2197 0:0057 þ j2:2568
6 7
¼ 4 0:0954 þ j2:2197 0:7374  j5:1374 0:1249 þ j2:8146 5
0:0057 þ j2:2568 0:1249 þ j2:8146 0:7236  j5:0289
5.3 An Example Three-Machine Power System 219

From Eq. (5.34), it can be obtained




¼ Y
Y  1 þ jx0
N d
2 3
0:8312  j3:0501 0:2806 þ j1:4829 0:2054 þ j1:2048
6 7
¼ 4 0:2806 þ j1:4829 0:4192  j2:7499 0:2119 þ j1:0795 5
0:2054 þ j1:2048 0:2119 þ j1:0795 0:2759  j2:3757
2 3
3:1614\1:3047 1:5092\1:3838 1:2221\1:4019
6 7
¼ 4 1:5092\1:3838 2:7816\1:4195 1:1001\1:3770 5
1:2221\1:4019 1:1001\1:3770 2:3917\1:4552

Thus according to Eq. (5.37), it can have


2 3 2 3
0:1450 0:1484 0:0034 2:8414 1:5907 1:2507
6 7 6 7
Fdd ¼ 4 1:4886 2:4386 0:9500 5; Fqq ¼ 4 0:6023 1:2064 0:6042 5;
1:1275 0:8120 1:9395 0:6478 0:8347 1:4825
2 3 2 3
3:0501 1:0626 0:9434 0:8312 1:0718 0:7770
6 7 6 7
Gdd ¼ 4 0:5975 2:7499 1:0467 5; Gqq ¼ 4 1:3859 0:4192 0:3384 5;
0:6326 1:0970 2:3757 1:0457 0:0823 0:2759
2 3 2 3
0:0300 0:7991 0:8366 0:1101 0:7923 1:0157
6 7 6 7
Hdd ¼ 4 0:0500 0:3126 0:3643 5; Hqq ¼ 4 0:0216 2:0503 1:1270 5:
0:0377 0:0614 0:2971 0:0228 0:8179 2:5579

From Eq. (5.38), following matrices are calculated


2 3 2 3
0:3252 0:0095 0:3157 2:1589 1:3566 0:8023
6 7 6 7
Fd ¼ 4 1:5477 2:4103 0:8626 5; Fq ¼ 4 0:2670 0:3235 0:0565 5;
1:1306 0:8935 2:0241 0:2296 0:1689 0:3985
2 3 2 3
2:2175 1:2518 1:1666 1:6110 0:8094 0:4956
6 7 6 7
Gd ¼ 4 0:1557 2:7758 0:9860 5; Gq ¼ 4 0:6318 0:1512 0:1474 5:
0:3995 1:1026 2:3982 0:4495 0:0527 0:1083
220 5 A Multi-machine Power System Installed …

Thus according to Eqs. (5.41) and (5.42), it can have


0
K1 ¼ ðVd  Iq  Xd )  Fd þ ðVq þ Id  Xq Þ  Fq ;
0
K2 ¼ Iq þ ðVd  Iq  Xd )  Gd þ ðVq þ Id  Xq Þ  Gq ;
0
K3 ¼ I þ (Xd  Xd )  Gd ;
0
K4 ¼ ðXd  Xd )  Fd ;
0
K5 ¼ V1 1
gen  Vq  Xd  Fd þ Vgen  Vd  Xq  Fq ;
0
K6 ¼ V1 1 1
gen  Vq  Vgen  Vd  Xd  Gd þ Vgen  Vd  Xq  Gq

Coefficient matrices of Heffron–Phillips model of example power system as


given by Eq. (5.43) are obtained to be
2 3 2 3
2:3373 1:4640 0:8733 2:4574 0:9034 0:5628
6 7 6 7
K1 ¼ 4 1:5290 2:2160 0:6870 5; K2 ¼ 4 1:0247 3:0671 0:4060 5;
1:0599 0:8201 1:8800 0:3694 0:6458 2:3234
2 3 2 3
1:1889 0:1067 0:0994 0:0277 0:0008 0:0269
6 7 6 7
K3 ¼ 4 0:1209 3:1565 0:7660 5; K4 ¼ 4 1:2024 1:8725 0:6701 5;
0:4521 1:2478 3:7141 1:2795 1:0112 2:2907
2 3 2 3
0:0068 0:0076 0:0143 0:8732 0:0711 0:0678
6 7 6 7
K5 ¼ 4 0:0596 0:0315 0:0281 5; K6 ¼ 4 0:4315 0:5292 0:1730 5:
0:0780 0:0499 0:1278 0:4691 0:1829 0:4790

2 3 2 3
47:2 0 0 0 0 0
6 7 6 7
M¼4 0 12:8 0 5; D ¼ 4 0 0 0 5;
0 0 6:02 0 0 0
2 3 2 3
200 0 0 0:02 0 0
6 7 6 7
KA ¼ 4 0 200 0 5; TA ¼ 4 0 0:02 0 5;
0 0 200 0 0 0:02
2 3
8:96 0 0
0 6 7
Td0 ¼4 0 6:00 0 5:
0 0 5:89
5.3 An Example Three-Machine Power System 221

5.3.2 Modal Analysis of Example Power System

5.3.2.1 Electromechanical Oscillation Modes of Example Power


System

Eigenvalues of state matrix of Heffron–Phillips model obtained in the above section


are calculated as

1;2 ¼ 25:0316  j28:2289


k
k3 ¼ 34:2648
4;5 ¼ 25:0343  j9:5156
k
6 ¼ 13:5354
k
7;8 ¼ 1:5051  j9:1860
k
9;10 ¼ 0:1736  j7:0590
k
11;12 ¼ 0:0000  j0:0000
k

Frequency of k 7;8 , and k


4;5 , k 9;10 is 9:5156 ¼ 1:5145 Hz, 9:1860 ¼ 1:4620 Hz, and
2p 2p
7:0590
2p ¼ 1:1235 Hz; respectively, falling within the range of power system
low-frequency oscillation 0.1–2 Hz. They can be identified to be the electrome-
chanical oscillation modes of the example power system. To confirm the identifi-
cation, participation factor and the correlation ratio of electromechanical loop can
be calculated for k 4 ¼ 25:0343 þ j9:5156, k 7 ¼ 1:5051 þ j9:1860, and k 9 ¼
0:1736 þ j7:0590 as follows.
Right and left eigenvectors corresponding to k 4 ¼ 25:0343 þ j9:5156 are
calculated as
2 3 2 3
0:6852  j17:3385 j0:0001
6 0:6595 þ j5:9033 7 6 0:0002 þ j0:0003 7
6 7 6 7
6 1:3447 þ j11:4352 7 6 0:0001 þ j0:0001 7
6 7 6 7
6 64:7503 þ j192:9710 7 6 0 7
6 7 6 7
6 31:8356  j61:9802 7 6 0 7
6 7 6 7
6 32:9147  130:9908 7 6 0 7
6
7 ¼ 6 7 6
7 ¼ 6 7
v
1:9362  396:4038 7; w 0:0020 þ j0:0008 7
6 7 6 7
6 2:5524 þ j88:6491 7 6 0:0044  j0:0017 7
6 7 6 7
6 2:5028 þ j50:7603 7 6 0:0024  j0:0010 7
6 7 6 7
6 0:5973  j1:5444 7 6 0:5202  j0:0006 7
6 7 6 7
4 0:18208 þ j0:5224 5 4 0:7505 þ j0:0000 5
0:12978 þ j0:2957 0:4076 þ j0:0030

By use of Eq. (5.60), participation factors corresponding to each of state vari-


ables are calculated as
222 5 A Multi-machine Power System Installed …

p41 ¼ 0:0017; p42 ¼ 0:0022; p43 ¼ 0:0018:

Thus, the correlation ratio of electromechanical loop can be calculated by using


Eq. (5.63) as r4 ¼ 0:0040, which is smaller than 1. Hence, k 4 ¼ 25:0343 þ
j9:5156 is not an electromechanical oscillation mode.
Right and left eigenvectors corresponding to k 7 ¼ 1:5051 þ j9:1860 are cal-
culated as
2 3 2 3
1:0451 þ j2:2184 0:0007  j0:0011
6 2:2567 þ j5:5000 7 6 0:0073  j0:0198 7
6 7 6 7
6 3:3018  j7:7184 7 6 0:0252 þ j0:0502 7
6 7 6 7
6 68:1811  j46:9144 7 6 0 7
6 7 6 7
6 170:8669  j105:1733 7 6 0:0005 þ j0:0003 7
6 7 6 7
6 239:0480 þ j152:08774 7 6 0:0013  j0:0010 7
7 ¼ 6
v 7; w
 ¼ 6
6 0:1000 þ j2:5611 7 7 6 0:0005  j0:0001 7
7
6 7 6 7
6 2:449 þ j6:2889 7 6 0:0039 þ j0:0037 7
6 7 6 7
6 3:5529  j8:2378 7 6 0:0042  j0:0188 7
6 7 6 7
6 0:0009 þ j0:0057 7 6 0:0017 þ j0:0447 7
6 7 6 7
4 0:0042 þ j0:0224 5 4 0:1917 þ j0:1403 5
0:0067  j0:0301 0:9683

By use of Eq. (5.60), participation factors corresponding to each of state vari-


ables are calculated as

p71 ¼ 0:1427; p72 ¼ 0:3021; p73 ¼ 0:0618:

Thus, the correlation ratio of electromechanical loop can be calculated by using


Eq. (5.63) as r7 ¼ 4:8776, which is greater than 1. Hence, k 7 ¼
1:5051 þ j9:1860 is an electromechanical oscillation mode.
9 ¼ 0:1736 þ j7:0590
Similarly, right and left eigenvectors corresponding to k
are calculated as
2 3 2 3
0:0830 þ j2:4518 0:0018  j0:0581
6 0:0192  j1:8299 7 6 0:0085 þ j0:1649 7
6 7 6 7
6 0:0638  j0:62186 7 6 0:0082 þ j0:0985 7
6 7 6 7
6 108:9599  j6:3716 7 6 0:0013 þ j0:0001 7
6 7 6 7
6 81:3708 þ j2:8540 7 6 0:0037  j0:0003 7
6 7 6 7
6 27:5890 þ j3:5176 7 6 0:0022  j0:0002 7
9 ¼ 6
v 7; w
 ¼ 6
6 0:6337 þ j0:5454 7 9 6 0:0015 þ j0:0016 7
7
6 7 6 7
6 0:9340  j0:2801 7 6 0:0066  j0:0195 7
6 7 6 7
6 0:2936  j0:0842 7 6 0:0030  j0:0134 7
6 7 6 7
6 0:0012 þ j0:0014 7 6 0:0973 þ j0:0918 7
6 7 6 7
4 0:0029  j0:0014 5 4 0:8031 5
0:0009  j0:0004 0:5441  j0:0031i
5.3 An Example Three-Machine Power System 223

By use of Eq. (5.60), participation factors corresponding to each of state vari-


ables are calculated as

p91 ¼ 0:0031; p92 ¼ 0:1254; p93 ¼ 0:4717:

Thus, the correlation ratio of electromechanical loop can be calculated by using


Eq. (5.63) as q9 ¼ 34:3291, which is greater than 1. Hence, k 9 ¼ 0:1736 þ
j7:0590 is also an electromechanical oscillation mode.
According to Eq. (5.66), mode shape of electromechanical oscillation mode can
be identified from the elements of right eigenvector corresponding to
Dxj ; j ¼ 1; 2; 3, which is the 4th, 5th, and 6th element of right eigenvector,
respectively. For k7 ¼ 1:5051 þ j9:1860, the elements of right eigenvector are

v74 ¼ 0:0000 þ j0:0000 ¼ 0\42:2017



75 ¼ 0:0005 þ j0:0003 ¼ 0:0006\29:5912
v
76 ¼ 0:0013  j0:0010 ¼ 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
4

G2
2

-2

-4

-6

-8

G3

-10
-12 -10 -8 -6 -4 -2 0 2 4 6
-4
x 10

7 ¼ 1:5051 þ j9:1860
Fig. 5.8 Mode shape of k
224 5 A Multi-machine Power System Installed …

x 10 -3
4

G1
0
G2 G3

-1

-2

-3

-4
-4 -3 -2 -1 0 1 2 3 4
-3
x 10

9 ¼ 0:1736 þ j7:0590
Fig. 5.9 Mode shape of k

9 ¼ 0:1736 þ j7.0590, the elements of right eigenvector are


Similarly, for k


v94 ¼ 0:0013 þ j0:0001 ¼ 0:0013\3:2273

v95 ¼ 0:0037  j0:0003 ¼ 0:0037\175:6366
96 ¼ 0:0022  j0:0002 ¼ 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 Heffron–Phillips model of example power system, the compact model of
system of Eq. (5.68) in the form of transfer function matrix is obtained as

Fðs) ¼ s2 M þ sD þ x0 K1  K2 ½ðI þ sTA ÞðK3 þ sT0d0 Þ þ KA K6 ]1


 ½ðI þ sTA ÞK4 þ KA Kk ]
5.3 An Example Three-Machine Power System 225

The Hooke-Jeeves optimum searching method in nonlinear programming can be


used to find the solution of following objective function of Eq. (5.74) to calculate
electromechanical oscillation modes as follows.
 
 ¼ DEL[FðkÞ
f(kÞ   ð5:74Þ

pffiffiffiffiffi
Step 1: Calculate eigenvalues of matrix x0 M1 K1 as gi ; i ¼ 1; 2. Thus, gi 6¼
0; j ¼ 1; 2 is the angular natural oscillation frequency of the electromechanical
oscillation modes. Choose k i ð0Þ ¼ 0 þ jpffiffiffiffi

gi on the complex plane as the starting
point of optimum searching.
Step 2: Choose an initial searching scale S0 [ 0, a minimum searching scale
SF [ 0, and an accelerating factor a [ 0. Search on the complex plane from k i ðk)

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

to ki ðk þ 1Þ are illustrated in Fig. 5.10 where S is the searching step and a ¼ 2. In
Fig. 5.10 if fðk  SÞ\fðk i Þ or fðk
  jSÞ\fðki Þ, the searching is successful in
i i
horizontal and vertical direction. Otherwise, it is a failure. From Fig. 5.10, it can be
seen that (1) pattern 1—there is one successful searching in both horizontal and

pattern 1 pattern 2

pattern 3

complex plane

the starting point λ i(k) the ending point λ i(k+1)


S 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 5 A Multi-machine Power System Installed …

vertical direction; (2) pattern 2—only one searching in the horizontal direction is
successful; (3) pattern 3—only 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;
Step 4: If at ki (kF ), the searching scale becomes less than SF , stop the searching and

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
The minimum searching scale SF [ 0 is the maximum distance of k i (kF ) to the

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 defines the accuracy of
optimum searching which is fully controllable. Of course, theoretically, the algo-
rithm 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 efficiency of an algorithm where one opera-
tion of an addition or multiplication is defined 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
inverse calculation of an n-order full matrix is n3 . Thus, the CC to form the
3

polynomial matrix model of Eq. (5.74) (i.e. forming all coefficient matrices in F(s))
is estimated as 2n3 =3 þ 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 þ KHF )16n3 =3, where kHS is the times of suc-
cessful searching and KHF the times of searching failure. KHF can be estimated as
kHF ¼ logðS0 =SF Þ= log R, where R is the ratio of reducing the searching scale when
the searching in one step fails. If R = 2 S0 ¼ 0:01; SF ¼ 0:000001, we can have
kHF 13. Hence, the total CC of the reduced-order method is about
102n3 þ kHS 16n3 =3 þ oðn2 Þ.
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
oscillation modes are k 1 ¼ 0:1736 þ j7:0590 and k 2 ¼ 1:5051 þ j9:1860,
which are exactly as same as those obtained in the previous section where they are
denoted as k 9 ¼ 0:1736 þ j7:0590 and k 7 ¼ 1:5051 þ j9:1860.
5.3 An Example Three-Machine Power System 227

Start

Choose i (0)=0+j i ,i=1,2 as the searching


Step 1 start points,where i are the eigenvalues of
matrix - 0 M -1K 1 and i 0.

Step 2 Initalization:i=k=m=1, d1 1 , d 2 j ,
set the value of S0 , SF and .

No Yes
f( i (k)+d mS)<f( i (k))?

No Yes
f( i (k)-d mS)<f( i (k))? i (k+1)= i (k)+d mS

i (k+1)= i (k) i (k+1)= i (k)-d mS

No Yes
m<2? m=m+1

No 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,
No Yes
i (k+1)= i (k), S<SF?
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 5 A Multi-machine Power System Installed …

SF

*
i

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 final point

7.6
Imaginary part 48000000
of eigenvalue
7.5 40000000
Start Point

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 0 0.05

1 calculation using direct searching algorithm


Fig. 5.13 Path of k
5.3 An Example Three-Machine Power System 229

11 Imaginary part
of eigenvalue
10.8 900000000

10.6 70000000
Start Point
0

10.4
500000
000
10.2

10
3000
9.8 0000
0

9.6 100000
000

9.4 50000
000

9.2
Final Point Real part
of eigenvalue
9
-1.6 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0

2 calculation using direct searching algorithm


Fig. 5.14 Path of k

5.3.2.3 Coordinated Design of PSSs

Damping ratio of k1 is 0.0246 and that of k


2 is 0.1617. Now PSSs will be installed
in three machines to enhance the damping ratio of k1 to 0.1. Parameters of PSSs can
be set by use of the method introduced previously in Sect. 5.2.2 to move k 1 to

k1 ¼ 0:7094 þ j7:0590. Hence, following objective function is set up
2 2
f ðpÞ ¼ ½Re k1 ðpÞ  Re k
 ðp Þ  þ ½Im k1 ðpÞ  Im k
 ðp Þ 
1 1

where p ¼ ½ Dpss1 Dpss2 Dpss3  and Dpssi is the coefficient of damping torque
provided by the PSS installed in the ith generator.
Choose p0 ¼ ½ 0 0 0  with f ðp0 Þ ¼ 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 ¼
½ 21:73 11:84 21:94  with f ðp52 Þ ¼ 9:2687  1015 , k 1 is moved to k c1 .
Figure 5.15 is the trajectories of Dpssi ; i ¼ 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
searching with respect to the movement of oscillation mode k 1 .
230 5 A Multi-machine Power System Installed …

25

20
Designed damping

15

10

5
Dpss1
Dpss2
Dpss3
0
0 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

Start Point
Objective function

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
5.3 An Example Three-Machine Power System 231

Table 5.6 Parameter of the Kpss T2 T4


PSSs
G1 4.1272 0.096 0.096
G2 1.4864 0.095 0.095
G3 3.0945 0.1 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

 Þ ¼ K2 KA ðK þ k
Fpss ðk  TA Þ þ K6 KA 1
 T0 ÞðI þ k
1 3 1 d0 1
2 3
4:0590\29:34 1:7719\153:70 1:0376\155:74
6 7
¼ 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

 Þ ¼ 4:0590\29:34
Fpss11 ðk 1
 Þ ¼ 6:1802\28:79
Fpss22 ðk 1
 Þ ¼ 5:3227\31:52
Fpss33 ðk 1

Let the transfer function of the PSS be

ð1 þ sT2i Þ ð1 þ sT4i Þ
Tpssi ðsÞ ¼ Kpssi ; T1i ¼ T3i ¼ 0:05
ð1 þ sT1i Þ ð1 þ sT3i Þ

By use of the phase compensation method for the PSS to compensate the for-
 Þ; i ¼ 1; 2; 3 to provide a pure damping torque
ward path of the ith generator Fii ðk

 i
Dpssi , where Dpss1 Dpss2 Dpss3 ¼ p52 ¼ ½ 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 ¼ Apss DX

where
 
Apss11 Apss12 Apss13
Apss ¼
Apss21 Apss22 Apss23
232 5 A Multi-machine Power System Installed …

2 3
0 0 0 314:1592 0 0
6 0 0 0 0 314:1592 0 7
6 7
6 314:1592 7
6 0 0 0 0 0 7
6 7
6 0:0495 0:0310 0:0185 0 0 0 7
6 7
Apss11 ¼6
6 0:1195 0:1731 0:0537 0 0 0 7
7
6 0:1761 0:3123 7
6 0:1362 0 0 0 7
6 7
6 0:0031 0:0001 0:0030 0 0 0 7
6 7
4 0:2004 0:3121 0:1117 0 0 0 5
0:2172 0:1717 0:3889 0 0 0

2 3
0 0 0 0 0 0
6 0 0 0 0 0 0 7
6 7
6 0 7
6 0 0 0 0 0 7
6 7
6 0:0521 0:0191 0:0119 0 0 0 7
6 7
Apss12 ¼6
6 0:0801 0:2396 0:0317 0 0 0 7 7
6 0:0614 0:1073 0:3860 0 7
6 0 0 7
6 7
6 0:1327 0:0119 0:0111 0:1116 0 0 7
6 7
4 0:0202 0:5261 0:1277 0 0:1667 0 5
0:0768 0:2119 0:6306 0 0 0:1698
2 3
0 0 0 0 0 0
60 0 0 0 0 07
6 7
60 07
6 0 0 0 0 7
6 7
60 0 0 0 0 07
6 7
Apss13 ¼6
60 0 0 0 0 077
60 07
6 0 0 0 0 7
6 7
60 0 0 0 0 07
6 7
40 0 0 0 0 05
0 0 0 0 0 0
2 3
67:5657 75:7691 143:3348 0 0 0
6 595:9666 314:8297 281:1369 0 0 0 7
6 7
6 779:5679 498:8968 1278:4646 0 7
6 0 0 7
6 7
6 0:0951 0:0596 0:0355 20 0 0 7
6 7
Apss21 ¼6
6 0:7703 0:4825 0:2878 162:048 0 0 7 7
6 0:2270 0:3289 0 7
6 0:1020 0 20 7
6 7
6 0:8193 1:1874 0:3681 0 72:2 0 7
6 7
4 0:3521 0:2724 0:6246 0 0 20 5
3:9227 3:0351 6:9579 0 0 222:8
5.3 An Example Three-Machine Power System 233

2 3
8731:9338 710:95188 678:1403 50 0 0
6 4315:4400 5291:9761 1729:6500 0 50 0 7
6 7
6 4691:1745 1828:9008 4789:5838 50 7
6 0 0 7
6 7
6 0:1000 0:0367 0:0229 0 0 0 7
6 7
Apss22 6
¼ 6 0:8099 0:2977 0:1855 0 0 0 77
6 0:1521 0:4553 0 7
6 0:0603 0 0 7
6 7
6 0:5491 1:6435 0:2176 0 0 0 7
6 7
4 0:1227 0:2146 0:7719 0 0 0 5
1:3673 2:3902 8:5990 0 0 0
2 3
0 10000 0 0 0 0
6 0 0 0 10000 0 0 7
6 7
6 10000 7
6 0 0 0 0 0 7
6 7
6 20 0 0 0 0 0 7
6 7
Apss23 ¼6
6 77:648 20 0 0 0 0 7 7
6 20 0 7
6 0 0 0 0 7
6 7
6 0 0 34:2 20 0 0 7
6 7
4 0 0 0 0 20 0 5
0 0 0 0 111:4 20

Eigenvalues of state matrix are calculated as

1 ¼ 44:7474 þ j0; k
k 2;3 ¼ 23:7299  j27:9946; k
4 ¼ 33:1534 þ j0,
5;6 ¼ 19:9770  j11:9203; k
k 7;8 ¼ 12:9863  j13:6705;
9;10 ¼ 0:6094  j7:0235; k
k 11;12 ¼ 4:5207  j8:0856;
13 ¼ 0:2772 þ j0,k
k 14 ¼ 0 þ j0, k
15 ¼ 23:6389 þ j0,

k 17 ¼ 17:7485 þ j0,k


16 ¼ 12:1456 þ j0,k 18 ¼ 15:9316 þ j0

Obviously, k 1 is moved to k c1 ¼ 0:6094 þ j7:0235, slightly away from the



 ¼ 0:7094 þ j7:0590. In Sect. 6.4.1, it is examined why
target required position k1
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, confirms the effec-
tiveness of PSSs designed by use of the introduced method.
234 5 A Multi-machine Power System Installed …

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 Mathematical Model of a Multi-machine Power


System Installed with Thyristor-Based FACTS
Stabilizers

6.1.1 Heffron–Phillips Model [1]

6.1.1.1 Heffron–Phillips Model of an N-Machine Power System


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 Y11 Y12 Vm
¼ ð6:1Þ
Ig Y21 Y22 Eg
 T  
where Ig ¼ Ig1 Ig2    IgN ; E g ¼ E  g2    E
 g1 E  m is the
 gN T ; V
vector of voltage at all non-internal nodes of generators in the power system and Eg

is defined by Eq. (5.34). “Deleting” all non-internal nodes is to remove Vm in the
network equations of Eq. (6.1) to have

Ig ¼ ðY  1 Y
 21 Y  22 ÞE
 12 þ Y g ð6:2Þ
11

where only N internal nodes of generators are kept. Comparing Eqs. (6.2) and (5.34),
it can be seen that in fact

 ¼Y
Y  1 Y
 21 Y  12 þ Y
 22 ð6:3Þ
11

© Springer Science+Business Media New York 2016 235


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_6
236 6 Multi-machine Power System Installed …

Fig. 6.1 Network of an E g1 Eg 2 E gN


N-machine power system Ig1 Ig 2 IgN
including internal nodes of
generators
x d1 ' xd2 ' x dN '

V g1 V g2 V gN

Transmission network YN

Fig. 6.2 An N-machine E1 E2 EN


power system installed with Ig1 Ig 2 IgN
an SVC-based stabilizer

X d1 ' Xd 2 ' X dN '

V g1 Vg 2 V gN

Transmission network 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  11
Y  12
Y  1n
V
Ig ¼  21  22 g ð6:4Þ
Y Y E

 11 ¼ y11 þ jbsvc and jbsvc is the equivalent admittance of the SVC.


where Y
6.1 Mathematical Model of a Multi-Machine Power System … 237

From Eq. (6.4), it can be obtained that

Ig ¼ 1  21 Y
 12 E
g þ Y g
 22 E
Y
ðy11 þ jbsvc Þ
  ð6:5Þ
1  21 Y
 12 þ Y
 22 Eg ¼ Y
Eg
¼ Y
ðy11 þ jbsvc Þ


From Eq. (6.5), it can be seen that the elements of network admittance matrix, Y,
can be denoted as
yij ¼ y0ij þ yij ðbsvc Þ ð6:6Þ

From Eqs. (5.35), (6.5), and (6.6), it can have

X
N
Igi ¼ idi þ jiqi ¼ yij ½E0qj ejð90

þ dj di Þ
þ ðxqj  x0dj Þiqj ejðdj di Þ 
j¼1

X
N

¼ y0ij ½E0qj ejð90 þ aij þ dj di Þ
þ ðxqj  x0dj Þiqj ejðaij þ dj di Þ  ð6:7Þ
j¼1

X
N

þ yij ðbsvc Þ½E0qj ejð90 þ dj di Þ
þ ðxqj  x0dj Þiqj ejðdj di Þ 
j¼1

0ij ¼ y0ij ejaij . From Eq. (6.7), it can be obtained that


where y
X
N
idi ¼ y0ij ½E0qj sinðaij  dij Þ þ ðxqj  x0dj Þiqj cosðaij  dij Þ
j¼1

X
N
þ Re½yij ðbsvc Þ½E0qj sin dij þ ðxqj  x0dj Þiqj cos dij 
j¼1

X
N
 Im½yij ðbsvc Þ½E0qj cos dij  ðxqj  x0dj Þiqj sin dij 
j¼1
ð6:8Þ
X
N
iqi ¼ yij ½E0qj cosðaij  dij Þ þ ðxqj  x0dj Þiqj sinðaij  dij Þ
j¼1

X
N
þ Re½yij ðbsvc Þ½E0qj cos dij  ðxqj  x0dj Þiqj sin dij 
j¼1

X
N
þ Im[yij ðbsvc Þ½E0qj sin dij þ ðxqj  x0dj Þiqj cos dij 
j¼1

where Re½yij ðbsvc Þ and Im½yij ðbsvc Þ denote the real and imaginary parts of yij ðbsvc Þ,
respectively.
238 6 Multi-machine Power System Installed …

Linearization of Eq. (6.8) is

DId ¼ Fdd Dd þ Gdd DE0q þ Hdd DIq þ Ldd Dbsvc


ð6:9Þ
DIq ¼ Fqq Dd þ Gqq DE0q þ Hqq DIq þ Lqq Dbsvc

The above equation can give

DId ¼ Fd Dd þ Gd DE0q þ Ld Dbsvc


ð6:10Þ
DIq ¼ Fq Dd þ Gq DE0q þ Lq Dbsvc

Substituting Eq. (6.10) into Eqs. (5.40) and (5.41), the Heffron–Phillips model of
the N-machine power system installed with the SVC stabilizer can be obtained as

Dd_ ¼ xo Dx
Dx_ ¼ M1 ðK1 Dd  K2 DE0q  DDx þ KP Dbsvc Þ
0 ð6:11Þ
DE_ q ¼ T01 0 0
d0 ðK3 DEq  K4 Dd þ DEfd þ Kq Dbsvc Þ
0
DE_ fd ¼ T1 0 1 0
A DEfd þ TA KA ðK5 Dd  K6 DEq þ KV Dbsvc Þ

6.1.1.2 Heffron–Phillips Model of an N-Machine Power System


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 
Y  12
Y  12
V
Ig ¼  11  22 g ð6:12Þ
Y21 Y E

where
2 3
y11 þ  1 y12  z 1
 11 4 z12 jxt csc 12  jxt csc 5
Y ¼
y21   1 y22 þ z 1jx
z12 jxt csc 12 t csc
 
 12
V ¼ V 1 V 2 T

Hence, the network admittance matrix with only the internal nodes of generators
left is
6.1 Mathematical Model of a Multi-Machine Power System … 239

Fig. 6.3 An N-machine E g1 Eg 2 E gN


power system installed with a Ig1 Ig 2 IgN
TCSC-based stabilizer

x d1 ' xd2 ' x dN '

V g1 Vg 2 V gN

Transmission network YN

y12 (y 21 )
V1 V2
Node 1 Node 2

z12
x t csc
y11 y 22

 ¼Y  1 Y
 21 Y  
Y 11 12 þ Y22 ð6:13Þ

Obviously, the element of the network admittance matrix yij can be denoted as

yij ¼ y0ij þ yij ðxt csc Þ ð6:14Þ

From Eqs. (5.35), (6.12), and (6.14), it can have


X
N
Igi ¼ idi þ jiqi ¼ yij ½E0qj ejð90

þ dj di Þ
þ ðxqj  x0dj Þiqj ejðdj di Þ 
j¼1

X
N

¼ y0ij ½E0qj ejð90 þ aij þ dj di Þ
þ ðxqj  x0dj Þiqj ejðaij þ dj di Þ  ð6:15Þ
j¼1

X
N

þ yij ðxt csc Þ½E0qj ejð90 þ dj di Þ
þ ðxqj  x0dj Þiqj ejðdj di Þ 
j¼1

Similar to Eqs. (6.7)–(6.10), by using Eq. (6.15), the Heffron–Phillips model of


the N-machine power system installed with the TCSC stabilizer can be obtained as
Dd_ ¼ xo Dx
Dx_ ¼ M1 ðK1 Dd  K2 DE0q  DDx þ KP Dxt csc Þ
0 ð6:16Þ
DE_ q ¼ T01 0 0
d0 ðK3 DEq  K4 Dd þ DEfd þ Kq Dxt csc Þ
0
DE_ fd ¼ T1 0 1 0
A DEfd þ TA KA ðK5 Dd  K6 DEq þ KV Dxt csc Þ
240 6 Multi-machine Power System Installed …

Fig. 6.4 An N-machine E g1 Eg 2 E gN


power system installed with a Ig1 Ig 2 IgN
TCPS-based stabilizer

x d1 ' xd2 ' x dN '

V g1 Vg 2 V gN

Transmission network 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 ¼ kej/ , it can have
    
0  11
Y  12
Y  12
V
Ig ¼  21  22 g ð6:17Þ
Y Y E

where
2 j/ 3
y11 þ 2
1 y12  e 
 11 ¼ 4 k z 12 kz 12
5
Y
y21  e
j/
y22 þ z112
kz12
 
 12n
V ¼ V1 V 2 T

Hence, the network admittance matrix with only the internal nodes of generators
left can be obtained as

 ¼Y  1 Y
 21 Y  
Y 11 12 þ Y22 ð6:18Þ

and the element of the network admittance matrix yij can be denoted as

yij ¼ y0ij þ yij ð/Þ ð6:19Þ


6.1 Mathematical Model of a Multi-Machine Power System … 241

Hence, for the system with the TCPS stabilizer installed, it can have

X
N
Igi ¼ idi þ jiqi ¼ yij ½E0qj ejð90

þ dj di Þ
þ ðxqj  x0dj Þiqj ejðdj di Þ 
j¼1

X
N

¼ y0ij ½E0qj ejð90 þ aij þ dj di Þ
þ ðxqj  x0dj Þiqj ejðaij þ dj di Þ  ð6:20Þ
j¼1

X
N

þ yij ð/Þ½E0qj ejð90 þ dj di Þ
þ ðxqj  x0dj Þiqj ejðdj di Þ 
j¼1

By taking the similar procedure from Eqs. (6.7) to (6.10), from Eq. (6.20), the
Heffron–Phillips model of the N-machine power system installed with the TCPS
stabilizer can be obtained as

Dd_ ¼ xo Dx
Dx_ ¼ M1 ðK1 Dd  K2 DE0q  DDx þ KP D/Þ
0 ð6:21Þ
DE_ q ¼ T01 0 0
d0 ðK3 DEq  K4 Dd þ DEfd þ Kq D/Þ
0
DE_ fd ¼ T1 0 1 0
A DEfd þ TA KA ðK5 Dd  K6 DEq þ KV D/Þ

The Heffron–Phillips model of the N-machine power system installed with an


SVC stabilizer, TCSC stabilizer, or TCPS stabilizer can be expressed in a unified
form as shown in Fig. 6.5. In Fig. 6.5, Dufactss ¼ Dbsvc for the SVC stabilizer,
Dufactss ¼ Dxt csc for the TCSC stabilizer, and Dufactss ¼ D/ for the TCPS sta-
bilizer, respectively.

K1

-
ω0 I
(sM + D) −1
s
- + Δu facts −s

KP Kq KV

K4 K5
K2

+ -
- + +
−1
(K 3 + sTd0 ) (I + sTA ) −1 K A Δu pss
+
-

K6

Fig. 6.5 Unified Heffron–Phillips model of an N-machine power system installed with a
thyristor-based FACTS stabilizer
242 6 Multi-machine Power System Installed …

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
Heffron–Phillips 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 ¼ Agxy Xg þ Bg Dupss þ Bgxy DVxy


ð6:22Þ
DIxy ¼ Cgxy Xg þ Dgxy DVxy

where

Agxy ¼ Ag þ Bgv BgV


Bgxy ¼ Bgv Tg0
Cgxy ¼ T1
g0 ðCg  BgI Þ
Dgxy ¼ 0

In the case that the reduced-order model of generators which is used for
establishing the Heffron–Phillips mode, from Eq. (2.31), it can have

vgdi E0  vgqi
iqi ¼ ; idi ¼ d 0 ð6:23Þ
xqi xd

Linearization of the above equation is


0
Dvgdi DEq  Dvgqi
Diqi ¼ ; Didi ¼ ð6:24Þ
xqi x0q
6.1 Mathematical Model of a Multi-Machine Power System … 243

From Eq. (6.24) and (5.39), it can be obtained that

DIdq ¼ E0 DE0iq þ Xdq1 DVdq


DP ¼ Vdq0 DIdq þ Idq0 DVdq
ð6:25Þ
DEq ¼ DE0q þ X0dd DIdq
DV ¼ V0dq DVdq

where

1
E0 ¼ feij g; eij ¼ ;
x0di
for i ¼ 1; 3; . . .; 2N  1; j ¼ 1; 2; . . .; N; otherwise eij ¼ 0
1
Xdq1 ¼ diagðxdqi Þ; xdqi ¼  0 ; for i ¼ 1; 3; . . .; 2N  1;
xdi
1
xdqi ¼  ¼ 0; for i ¼ 2; 4; . . .2N
xqi
Vdq0 ¼ fVdq0ij g; Vdq0ij ¼ vgdi0 ; Vdq0ij þ 1 ¼ vgqi0 ;
for i ¼ 1; 2; . . .; N; j ¼ 1; 3; . . .; 2N  1; otherwise Vdq0ij ¼ 0
Idq0 ¼ fIdq0ij g; Idq0ij ¼ igdi0 ; Idq0ij þ 1 ¼ igqi0 ;
for i ¼ 1; 2; . . .; N; j ¼ 1; 3; . . .; 2N  1; otherwise Idq0ij ¼ 0
X0dd ¼ fx0ddij g; x0ddij ¼ ðxdi  x0di Þ;
for i ¼ 1; 2; . . .; N; j ¼ 1; 3; . . .; 2N  1; otherwise x0ddij ¼ 0
vgdi0 vgqi0
V0dq ¼ fV0dqij g; V0dqij ¼ ; V0dqij þ 1 ¼ ;
Vgi0 Vgi0
for i ¼ 1; 2; . . .; N; j ¼ 1; 3; . . .; 2N  1; otherwise Vdq0ij ¼ 0

By substituting the above equation into (5.40), it can be obtained that

Dd_ ¼ xo Dx
Dx_ ¼ M1 ðDDx  Vdq0 E0 DE0iq Þ  M1 ðVdq0 Xdq1 þ Idq0 ÞDVdq
0
DE_ q ¼ T01 0 0 0 0 01 0
d0 ðDEq  Xdd E0 DEiq þ DEfd Þ  Td0 Xdd Xdq1 DVdq ð6:26Þ
0
DE_ fd ¼ T1 0 1 1
A DEfd  TA KA V0dq DVdq þ TA KA Dupss
DIdq ¼ E0 DE0iq þ Xdq1 DVdq
244 6 Multi-machine Power System Installed …

which can be written as


2 3 2 32 3
Dd_ 0 xo I 0 0 Dd
6 7 6
6 Dx_ 7 6 0 M1 D M1 Vdq0 E0 0 7 6
76 Dx 7
7
6 0 7¼6 7 6 7
6 DE_ 7 4 0 T1 0
T1 54 DE0q 5
4 q 5 0 d0 ðI þ Xdd E0 Þ d0
0
DE_ fd 0 0 0 T1A DE0fd
2 3 2 3
0 0
6 M1 ðV X þ I Þ 7 6 0 7
6 dq0 dq1 dq0 7 6 7 ð6:27Þ
6 7DVdq þ 6 7Dupss
4 T01
d0 X 0
X
dd dq1
5 4 0 5
T1
A KA V0dq T1
A KA
2 3
Dd
6 Dx 7
6 7
DIdq ¼ ½ 0 0 E0 0 6 7 þ Xdq1 DVdq
4 DE0q 5
DE0fd

Obviously, the above equation can be written in the similar form to Eq. (5.48) as

X_ g ¼ Ag Xg þ Bg Dupss þ Bgv DVdq


ð6:28Þ
DIdq ¼ Cg Xg þ Dg DVdq

where
2 3 2 3
Dd 0 xo I 0 0
6 Dx 7 6 0 M1 D 1
M Vdq0 E0 7
6 7 6 0 7
Xg ¼ 6 7; Ag ¼ 6 7;
4 DE0q 5 40 0 T1 0
d0 ðI þ Xdd E0 Þ T1
d0
5
DE0 0 0 0 T1
2 fd 3 2 3 A
0 0
6 7 6 7
6 M1 ðVdq0 Xdq1 þ Idq0 Þ 7 6 0 7
6 7 6 7
Bgv ¼ 6 7; Bg ¼ 6 7;
6 T 01 0
X X 7 6 0 7
4 d0 dd dq1 5 4 5
T1
A KA V0dq T1
A KA

Cg ¼ ½ 0 0 E0 0 ; Dg ¼ Xdq1
By substituting Eq. (5.51) into the above equation, it can be obtained that

X_ g ¼ Agxy Xg þ Bg Dupss þ Bgxy DVxy


ð6:29Þ
DIxy ¼ Cgxy Xg þ Dgxy DVxy
6.1 Mathematical Model of a Multi-Machine Power System … 245

where

Agxy ¼ Ag þ Bgv BgV


Bgxy ¼ Bgv Tg0
Cgxy ¼ T1
g0 ðCg  BgI Þ

Dgxy ¼ 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 ðsÞ and the feedback signal Dxi , that is

Dupssi ¼ Gi ðsÞDxi ð6:30Þ

If the state-space realization of Eq. (6.30) is

X_ pssi ¼ Apssi Xpssi þ Bpssi Dxi


ð6:31Þ
Dupssi ¼ Cpssi Xpssi þ Dpssi Dxi

State-space representation of L PSSs installed in the N-machine power system


can be written as

X_ pss ¼ Apss Xpss þ Bpss Dx


ð6:32Þ
Dupss ¼ Cpss Xpss þ Dpss Dx

where
2 3 2 3
Xpss1 Bpss1 0 ... 0
6 Xpss2 7 6 0 Bpss2 ... 0 7
6 7 6 7
Xpss ¼6 7 6
6 .. 7; Bpss ¼ 6 .. .. ..
7;
7
4 . 5 4 . . ... . 5
XpssL 0 0 ... BpssL
2 3 2 3
Apss1 0 ... 0 Cpss1 0 ... 0
6 0 Apss2 ... 0 7 6 0 Cpss2 ... 0 7
6 7 6 7
Apss ¼6
6 .. .. ..
7; Cpss
7 ¼6
6 .. .. ..
7;
7
4 . . ... . 5 4 . . ... . 5
0 0 ... ApssL 0 0 ... CpssL
2 3
Dpss1 0 ... 0
6 0 Dpss2 ... 0 7
6 7
Dpss ¼6
6 .. .. .. 7
7
4 . . ... . 5
0 0 ... DpssL
246 6 Multi-machine Power System Installed …

By substituting Eq. (6.32) into Eqs. (6.22) or (6.29), it can be obtained that

X_ gp ¼ Agp Xgp þ Bgp DVxy


ð6:33Þ
DIxy ¼ Cgp Xgp þ Dgp DVxy

where
   
Xg Agxy þ Bg DDpss Bg Cpss
Xgp ¼ ; Agp ¼ ;
Xpss BDpss Apss
   
Bgxy Dgp
Bgp ¼ ; Cgp ¼ ½ Cgxy 0 ; Dgp ¼ ;
0 0
2 3 2 3
0 0 0 0 0 0
6 7 6 7
DDpss ¼ 4 0 Dpss 0 5; BDpss ¼ 4 0 Bpss 0 5
0 0 0 0 0 0

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 fixed 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
1 2ai  sin 2ai
bsvci ¼  ð6:34Þ
xsvcci 2pxsvcli

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 ðsÞ and Tssvci ðsÞ, 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 Vi Vj


the ith node in the N-machine Pij zij Iij
power system

Ii jbsvci
6.1 Mathematical Model of a Multi-Machine Power System … 247

ai ¼ avi þ asi ð6:35Þ

where
avi ¼ Tvsvci ðsÞðVi  Viref Þ
1 ð6:36Þ
asi ¼ Tssvci ðsÞ ðPij  Pijref Þ
s

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

1  cos 2ai0
Dbsvci ¼  ðDavi þ Dasi Þ ð6:37Þ
pxsvcli

Davi ¼ Tvsvci ðsÞDVi


1 ð6:38Þ
Dasi ¼ Tssvci ðsÞ DPij
s

where
vix0 viy0 h i Dv  h i
viy0 viy0
DVi ¼ Dvix þ Dviy ¼ vVix0 ix
¼ vVix0 Vi0 DVixy
Vi0 Vi0 i0 Vi0 Dviy i0

From Fig. 6.6, it can have


i  V
V j
Pij ¼ Reð   Þ ¼ Re½ðgij þ jbij ÞðV
V   V
 iV   Þ
 jV
zij i i i
ð6:39Þ
¼ gij ðv2xi þ v2yi Þ þ bij ðvxi vyi  vyi vxj Þ

where * denotes the conjugate of a phasor. Linearization of the above equation


gives
DPij ¼ 2gij ðvxi0 Dvxi þ vyi0 Dvyi Þ þ bij ðvyj0 Dvxi  vxj0 Dvyi Þ
þ bij ðvxi0 Dvyj  vyi0 Dvxj Þ
¼ ð2gij vxi0 þ bij vyj0 ÞDvxi þ ð2gij vyi0  bij vxj0 ÞDvyi
þ bij ðvxi0 Dvyj  vyj0 Dvxj Þ
 
  Dvxi
¼ 2g v
ij xi0 þ b v
ij yj0 2g v
ij yi0  b v
ij xj0
Dvyi
 
Dvxj
þ ½ bij vxi0 bij vyi0 
Dvyj
 
¼ 2gij vxi0 þ bij vyj0 2gij vyi0  bij vxj0 DVxyi
þ ½ bij vxi0 bij vyi0 DVxyj
¼ psvci DVxyi þ p0svcj DVxyj ð6:40Þ
248 6 Multi-machine Power System Installed …

Hence, from Eqs. (6.38) and (6.40), it can have


h i
v
Davi ¼ Tvsvci ðsÞ vVxi0 Vyi0 DVxyi
i0 i0

1 ð6:41Þ
Dasi ¼ Tssvci ðsÞ ðpsvci DVxyi þ p0svcj DVxyj Þ
s

Let the state-space realization of the voltage controller and stabilizer of the SVC
given by Eq. (6.41) be

X_ VSVCi ¼ AVSVCi XVSVCi þ BVSVCi DVxyi


ð6:42Þ
Davi ¼ CVSVCi XVSVCi þ DVSVCi DVxyi

X_ SSVCi ¼ ASSVCi XSSVCi þ BSSVCi DVxyi þ BSSVCj DVxyj


ð6:43Þ
Dasi ¼ CSSVCi XSSVCi þ DSSVCi DVxyi þ DSSVCj DVxyj

According to Fig. 6.6, the injected current from the SVC into the ith node is

Ii ¼ ixi þ jiyi ¼ jbsvci V


 i ¼ jbsvci ðvxi þ jvyi Þ ð6:44Þ

The above equation gives

ixi ¼ bsvci vyi ; iyi ¼ bsvci vxi ð6:45Þ

By using Eq. (6.37), linearization of Eq. (6.45) can be obtained as


      
Dixi vyi0 bsvci0 0 Dvxi
¼ Dbsvci þ
Diyi vxi0 0 bsvci0 Dvyi
   
1  cos 2ai0 vyi0 bsvci0 0
¼ ðDavi þ Dasi Þ þ DVxyi
pXsvcli vxi0 0 bsvci0
ð6:46Þ

When only the SVC voltage control function is considered, by substituting


Eq. (6.42) into (6.46), it can have
      
Dixi vyi0 bsvci0 0 Dvxi
¼ Dbsvci þ
Diyi vxi0 0 bsvci0 Dvyi
 
1  cos 2ai0 vyi0
¼ ðCVSVCi XVSVCi þ DVSVCi DVxyi þ DaSi Þ
pxsvcli vxi0
 
bsvci0 0
þ DVxyi ð6:47Þ
0 bsvci0
6.1 Mathematical Model of a Multi-Machine Power System … 249

That is

DISVCi ¼ CSVCi XVSVCi þ DSVCi DVxyi þ bSSVCi Dasi ð6:48Þ

where
   
Dixi 1  cos 2ai0 vyi0
DISVCi ¼ ; CSVCi ¼  CVSVCi ;
Diyi pxsvcli vxi0
   
1  cos 2ai0 vyi0 1  cos 2ai0 vyi0
DSVCi ¼ DVSVCi ; bSSVCi ¼ 
pxsvcli vxi0 pxsvcli vxi0

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 1  cos 2ai0 vyi0
¼ ðCVSVCi XVSVCi þ CSSVCi XSSVCi Þ
Diyi pxsvcli vxi0
 
1  cos 2ai0 vyi0
 ðDVSVCi DVxyi þ DSSVCi DVxyi þ DSSVCj DVxyj Þ
pxsvcli vxi0
 
bsvci0 0
þ DVxyi
0 bsvci0
ð6:49Þ

Put the state equation of the SVC voltage controller of Eq. (6.42) and stabilizer
of Eq. (6.43) together

X_ SVCi ¼ ASVCi XSVCi þ BSVCi DVxyi þ BSVCj DVxyj ð6:50Þ

where
   
XVSVCi AVSVCi 0
XSVCi ¼ ; ASVCi ¼ ;
XSSVCi 0 ASSVCi
   
BVSVCi 0
BSVCi ¼ ; BSVCj ¼
BSSVCi BSSVCj

Equation (6.49) can be written as

DISVCi ¼ CSVCi XSVCi þ DSVCi DVxyi þ DSVCj DVxyj ð6:51Þ


250 6 Multi-machine Power System Installed …

where
 
Dixi
DISVCi ¼
Diyi
     
vyi0 vyi0
CSVCi ¼  1cos 2ai0
pxsvcli CVSVCi  1cos
CSSVCi 2ai0
pxsvcli
vxi0 vxi0
   
1  cos 2ai0 vyi0 bsvci0 0
DSVCi ¼  ðDVSVCi þ DSSVCi Þ þ
pxsvcli vxi0 0 bsvci0
 
1  cos 2ai0 vyi0
DSVCj ¼  DSSVCj
pxsvcli vxi0

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 ¼ Vj  Vi ; Ij ¼ Vi  Vj ð6:52Þ
jxtcsci jxtcsci

Vi Vj
Pij
Vi Vj

Ii Ij Ii Ij

jx tcsci

Fig. 6.7 Creation of two extra nodes at the location where the TCSC is installed
6.1 Mathematical Model of a Multi-Machine Power System … 251

That is

Ii ¼ ixi þ jiyi ¼  j ðvxj þ jvyj  vxi  jvyi Þ


xtcsci
ð6:53Þ
Ij ¼ ixj þ jiyj ¼  j ðvxi þ jvyi  vxj  jvyj Þ
xtcsci

The above equation gives

1 1
ixi ¼ ðvyj  vyi Þ; iyi ¼ ðvxi  vxj Þ;
xtcsci xt csci
ð6:54Þ
1 1
ixj ¼ ðvyi  vyj Þ; iyj ¼ ðvxj  vxi Þ
xtcsci xtcsci

The normal control function of the TCSC is the load flow regulation. Let the
transfer function of the TCSC load flow regulator be Tltcsci ðsÞ. It can be assumed
that the transfer function of the TCSC stabilizer and feedback signal is Tstcsci ðsÞ 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 ¼ ½Tltcsci ðsÞ þ Tstcsci ðsÞðPij  Pijref Þ ð6:55Þ
s

where Pijref is the reference signal of the TCSC load flow and stabilizing control.
Linearization of Eq. (6.55) is

1
Dxtcsci ¼ ½Tltcsci ðsÞ þ Tstcsci ðsÞDPij ð6:56Þ
s

From Fig. 6.7, it can have

i  V
V j 
Pij ¼ Reð  Þ ¼ Re½ 1 ðV
V   V
 iV   Þ
 jV
i i i
jxtcsci jxtcsci
ð6:57Þ
1
¼ ðvyi vxj  vxi vyj Þ
xtcsci

Linearization of the above equation is

1
DPij ¼  ðvyi0 vxj0  vxi0 vyj0 ÞDvtcsci
x2tcsci0
1 1 ð6:58Þ
þ ðvxj0 Dvyi  vyj0 Dvxi Þ þ ðvyj0 Dvxj  vxi0 Dvyj Þ
xtcsci0 xtcsci0
¼ atcsc0 Dxtcsci þ atcsci DVxyi þ atcscj DVxyj
252 6 Multi-machine Power System Installed …

Fig. 6.8 TCSC with a PI load


K ptcsc
flow controller superimposed
by a stabilizing signal
ΔPij
+ + Δx tcsci

K itcsc Δα stcsci
s Δx ltcsci

where

1
atcsc0 ¼ ðvyi0 vxj0  vxi0 vyj0 Þ
x2tcsci0
h i
atcsci ¼  xtcsci0
1 1
vyj0 xtcsci0 vxj0
h i
atcscj ¼ xtcsci0
1
vyj0  xtcsci0
1
vxi0

For the case that the TCSC stabilizer is not installed, Eq. (6.56) becomes

Dxtcsci ¼ Tltcsci ðsÞDPij þ Dastcsci ð6:59Þ

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 ¼ Dastcsci þ Kptcsc DPij þ Dxltcsci


ð6:60Þ
Dx_ ltcsci ¼ Kitcsc DPij

From Eqs. (6.58) and (6.60), it can have

1 1
Dxtcsci ¼ Dastcsci þ Dxltcsci
1 þ Kptcsc atcsc0 1 þ Kptcsc atcsc0
ð6:61Þ
Kptcsc
þ ðatcsci DVxyi þ atcscj DVxyj Þ
1 þ Kptcsc atcsc0

Kitcsc atcsc0 Kitcsc atcsc0


Dx_ ltcsci ¼  Dxltcsci  Dastcsci
1 þ Kptcsc atcsc0 1 þ Kptcsc atcsc0
ð6:62Þ
Kptcsc Kitcsc atcsc0
þ ðKitcsc  Þðatcsci DVxyi þ atcscj DVxyj Þ
1 þ Kptcsc atcsc0

Equation (6.62) can be written as

Dx_ ltcsci ¼ Altcsci Dxltcsci þ BLTCSCi DVxyi þ BLTCSCj DVxyj þ bltcsci Dastcsci ð6:63Þ
6.1 Mathematical Model of a Multi-Machine Power System … 253

Linearization of Eq. (6.54) is


1 1
Dixi ¼  ðvyj0  vyi0 ÞDxtcsci þ
ðDvyj  Dvyi Þ;
x2tcsci0 xtcsci0
1 1
Diyi ¼  2 ðvxi0  vxj0 ÞDxtcsci þ ðDvxi  Dvxj Þ;
xtcsci0 xtcsci0
ð6:64Þ
1 1
Dixj ¼  2 ðvyi0  vyj0 ÞDxtcsci þ ðDvyi  Dvyj Þ;
xtcsci0 xtcsci0
1 1
Diyj ¼  2 ðvxj0  vxi0 ÞDxtcsci þ ðDvxj  Dvxi Þ
xtcsci0 xtcsci0

That is
2 3
2 3  x21 ðvyj0  vyi0 Þ
Dixi 6 tcsci0
7
6 Di 7 6  1
ðvxi0  vxj0 Þ 7
6 yi 7 6 x 2 7
6 7¼6 7Dxtcsci
tcsci0

4 Dixj 5 6 7
6  x2tcsci0 ðvyi0  vyj0 Þ 7
1
4 5
Diyj  x21 ðvxj0  vxi0 Þ ð6:65Þ
2 32 3
tcsci0

0 1 0 1 Dvxi
1 6 6 1 0 1 0 7 6 7
76 Dvyi 7
þ 6 76 7
xtcsci0 4 0 1 0 1 54 Dvxj 5
1 0 1 0 Dvyj

By using Eq. (6.61), the above equation becomes


DITCSCij ¼ CTCSCij Dxltcsci þ bTCSCij Dastcsci þ DTCSCij DVxyij ð6:66Þ

where
2 3 2 3
Dixi Dvxi
6 Diyi 7 6 Dvyi 7
6 7 6 7
DITCSCij ¼6 7; DVxyij ¼ 6 7;
4 Dixj 5 4 Dvxj 5
Diyj Dvyj
2 3
 x21 ðvyj0  vyi0 Þ
6 tcsci0
7
6  21 ðvxi0  vxj0 Þ 7
1 6 xtcsci0 7
bTCSCij ¼ CTCSCij ¼ 6 7;
1 þ Kptcsc atcsc0 6 7
6  x2tcsci0 ðvyi0  vyj0 Þ 7
1
4 5
 x21 ðvxj0  vxi0 Þ
2 3
tcsci0

0 1 0 1
 
Kptcsc atcsci 0 1 6 6 1 0 1 0 77
DTCSCij ¼ þ 6 7 ð6:67Þ
1 þ Kptcsc atcsc0 0 atcscj x tcsci0 4 0 1 0 1 5
1 0 1 0
254 6 Multi-machine Power System Installed …

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 ¼ dtcsci Dastcsci þ ctcsci Dxltcsci þ ptcsci DVxyi þ p0tcscj DVxyj ð6:68Þ

where
atcsc0 atcsc0
dtcsci ¼ ; ctcsci ¼ ;
1 þ Kptcsc atcsc0 1 þ Kptcsc atcsc 0
   
atcsc0 Kptcsc 0 atcsc0 Kptcsc
ptcsci ¼ 1 atcsci ; ptcscj ¼ 1  atcscj
1 þ Kpt csc atcsc0 1 þ Kptcsc atcsc0

When the TCSC stabilizer is installed as shown in Eq. (6.55), denote


Ttcsci ðsÞ ¼ Tltcsci ðsÞ þ s1 Tstcsci ðsÞ. From Eqs. (6.56) and (6.58), it can have

Ttcsci ðsÞ
Dxtcsci ¼ ðatcsci DVxyi þ atcscj DVxyj Þ
1 þ atcsc0 Ttcsci ðsÞ ð6:69Þ
¼ T0t csc i ðsÞðatcsci DVxyi þ atcscj DVxyj Þ

Let the state-space realization of Eq. (6.69) be

X_ TCSCi ¼ ATCSCi XTCSCi þ BTCSCi DVxyi þ BTCSCj DVxyj ð6:70Þ

Dxtcsci ¼ CTCSCi XTCSCi þ DTCSCi DVxyi þ DTCSCj DVxyj ð6:71Þ

By substituting Eq. (6.71) into Eq. (6.65), it can be obtained that

DITCSCij ¼ CTCSCij XTCSCi þ DTCSCij DVxyij ð6:72Þ

where
2 3
2 3  x21 ðvyj0  vyi0 Þ
Dixi 6 tcsci0
7
6 Di 7 6  21 ðvxi0  vxj0 Þ 7
6 7 6 x 7
7; CTCSCij ¼ 6 7
yi
DITCSCij ¼6 6  1 ðv  v Þ 7CTCSCi ;
tcsci0

4 Dixj 5 6 x2tcsci0 yi0 yj0 7


4 5
Diyj  x21 ðvxj0  vxi0 Þ
2 3 2 3
tcsci0

Dixi 0 1 0 1
6 Di 7   6 1
6 yi 7 D TCSCi 0 1 6 0 1 0 7
7
DVxyij ¼6 7; DTCSCij ¼ þ 6 7
4 Dixj 5 0 DTCSCj xtcsci0 4 0 1 0 1 5
Diyj 1 0 1 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 Mathematical Model of a Multi-Machine Power System … 255

6.1.2.4 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
 i ¼ I0 ¼ e ð6:73Þ
i j/i
V ij

0   j/i   
I0 ¼ Vi  Vj ¼ Vi e  Vi þ Vi  Vj
ij
zij zij
ð6:74Þ
  j/i i  V
j þ V
j  V
 j ej/i
Iij ¼ Iij ej/i ¼ Vi  Vj e ¼
V
zij zij

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 ¼ Vi e  Vi
zij
ð6:75Þ
 j/i j
V
Ii ¼ Vj e
zij

Denote z1ij ¼ gij þ jbij . From Eq. (6.75), it can be obtained that

ixi ¼ ðgij vxj þ bij vyj Þðcos/i 1Þ þ ðgij vyj þ bij vxj Þsin/i
iyi ¼ ðgij vxj þ bij vyj Þsin/i þ ðgij vyj þ bij vxj Þðcos/i 1Þ
ð6:76Þ
ixj ¼ ðgij vxi þ bij vyi Þðcos/i  1Þ  ðgij vyi þ bij vxi Þsin/i
iyj ¼ ðgij vxi þ bij vyi Þsin/i þ ðgij vyi þ bij vxi Þðcos/i 1Þ

Vi Vj
z iijj
Pij
Vi Vj
I ij I ij ' z iijj

Vi '
k :1
Ii Ij
k =1

Fig. 6.9 Installation of a TCPS between node i and j in an N-machine power system
256 6 Multi-machine Power System Installed …

Linearization of the above equation is

Dixi ¼ ½ðgij vxj0 þ bij vyj0 Þ sin /i0 þ ðgij vyj0 þ bij vxj0 Þ cos /i0 D/i
þ ½gij ðcos /i0  1Þ þ bij sin /i0 Dvxj þ ½bij ðcos /i0  1Þ þ gij sin /i0 Dvyj
Diyi ¼ ½ðgij vxj0 þ bij vyj0 Þ cos /i0   ðgij vyj0 þ bij vxj0 Þ sin /i0 þ D/i
þ ½gij sin /i0 þ bij ðcos /i0  1ÞDvxj þ ½bij sin /i0 þ gij ðcos /i0  1ÞDvyj
Dijx ¼ ½ðgij vxi0 þ bij vyi0 Þ sin /i0  ðgij vyi0 þ bij vxi0 Þ cos /i0 D/i
þ ½gij ðcos /i0  1Þ  bij sin /i0 Dvxi þ ½bij ðcos /i0  1Þ  gij sin /i0 Dvyi
Dijy ¼ ½ðgij vxi0 þ bij vyi0 Þ cos /i0  ðgij vyi0 þ bij vxi0 Þ sin /i0 D/i
þ ½gij sin /i0 þ bij ðcos /i0  1ÞDvxi þ ½bij sin /i0 þ gij ðcos /i0  1ÞDvyi

That is

DITCPSij ¼ bTCPSi D/i þ DTCPSi DVxyij ð6:77Þ

where
2 3 2 3
Dixi Dvxi
6 Di 7 6 Dv 7
6 yi 7 6 yi 7
DITCPSij ¼6 7; DVxyij ¼ 6 7;
4 Dixj 5 4 Dvxj 5
Diyj Dvyj
2 3
ðgij vxj0 þ bij vyj0 Þ sin /i0 þ ðgij vyj0 þ bij vxj0 Þ cos /i0
6 ðg vxj0 þ bij vyj0 Þ cos /  ðg vyj0 þ bij vxj0 Þ sin / 7
6 ij i0 ij i0 7
bTCPSi ¼6 7;
4 ðgij vxi0 þ bij vyi0 Þ sin /i0  ðgij vyi0 þ bij vxi0 Þ cos /i0 5
ðgij vxi0 þ bij vyi0 Þ cos /i0  ðgij vyi0 þ bij vxi0 Þ sin /i0
 
0 DTCSCij
DTCPSi ¼
DTCSCii 0
" #
gij ðcos /i0  1Þ þ bij sin /i0 bij ðcos /i0  1Þ þ gij sin /i0
DiTCSCj ¼
gij sin /i0 þ bij ðcos /i0  1Þ bij sin /i0 þ gij ðcos /i0  1Þ
" #
gij ðcos /i0  1Þ  bij sin /i0 bij ðcos /i0  1Þ  gij sin /i0
DiTCSCi ¼
gij sin /i0 þ bij ðcos /i0  1Þ bij sin /i0 þ gij ðcos /i0  1Þ

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
6.1 Mathematical Model of a Multi-Machine Power System … 257

Fig. 6.10 A PI TCPS load


K ptcps
flow controller superimposed
with a stabilizing signal
ΔPij
+ + Δφi

K itcps Δα stcpsi
s Δx ltcpsi

D/i ¼ Dastcpsi þ Kpt c ps DPij þ Dxltcpsi


ð6:78Þ
Dx_ ltcpsi ¼ Kit c ps DPij

From Eq. (6.74), it can have

  j/i
  Vi  Vj e
 i  Iij ¼ V
Sij ¼ V i z ij
  j ej/i Þ ð6:79Þ
¼ ðgij þ jbij ÞðVi  Vi  V
2

¼ ðgij þ jbij ÞV2i  ðgij þ jbij Þðvxi  jvyi Þðvxj þ jvyj Þðcos /i  j sin /i Þ

Hence,

Pij ¼ ReðSij Þ ¼ gij ðv2xi þ v2yi Þ


þ gij ½ðvxi vxj þ vyi vyj Þ cos /i þ ðvxi vyj  vyi vxj Þ sin /i  ð6:80Þ
 bij ½ðvxi vxj þ vyi vyj Þ sin /i þ ðvxi vyj  vyi vxj Þ cos /i 

Linearization of Eq. (6.80) is

DPij ¼ at c ps0 D/i þ atcpsi DVxyi þ atcpsj DVxyj ð6:81Þ

where

at c ps0 ¼ gij ½ðvxi0 vxj0 þ vyi0 vyj0 Þ sin /i0  ðvxi0 vyj0  vyi0 vxj0 Þ cos /i0 
þ bij ½ðvxi0 vxj0 þ vyi0 vyj0 Þ cos /i0 þ ðvxi0 vyj0  vyi0 vxj0 Þ sin /i0 
" #T
2gij vxi0 þ gij ðvxj0 cos /i0 þ vyj0 sin /i0 Þ  bij ðvxj0 sin /i0 þ vyj0 cos /i0 Þ
atcpsi ¼
2gij vyi0 þ gij ðvyj0 cos /i0  vxj0 sin /i0 Þ þ bij ðvyj0 sin /i0 þ vxj0 cos /i0 Þ
" #T
gij ðvxi0 cos /i0  vyi0 sin /i0 Þ þ bij ðvxi0 sin /i0 þ vyi0 cos /i0 Þ
atcpsj ¼
gij ðvyi0 cos /i0 þ vxi0 sin /i0 Þ  bij ðvyi0 sin /i0 þ vxi0 cos /i0 Þ
258 6 Multi-machine Power System Installed …

From Eqs. (6.78) and (6.81), it can have

1
D/i ¼ Dastcpsi
1 þ atcps0 Kptcps
Kptcps 1
þ ðatcpsi DVxyi þ atcpsj DVxyj Þ þ Dxltcpsi
1 þ atcps0 Kptcps 1 þ atcps0 Kptcps
ð6:82Þ

atcps0 Kitcps atcps0 Kitcps


Dx_ ltcpsi ¼ Dxltcpsi þ Dastcpsi
1 þ atcps0 Kptcps 1 þ atcps0 Kptcps
ð6:83Þ
atcps0 Kitcps Kptcps
þ ðKptcps  ðatcpsi DVxyi þ atcpsj DVxyj Þ
1 þ atcps0 Kitcps

Equation (6.83) can be written as

Dx_ ltcpsi ¼ Altcpsi Dxltcpsi þ BLTCPSi DVxyi þ BLTCPSj DVxyj þ bltcpsi Dastcpsi ð6:84Þ

From Eqs. (6.77) and (6.82), it can have

DITCPSij ¼ CTCPSij xltcpsi þ bTCPSij Dastcpsi þ DTCPSij DVxyij ð6:85Þ

where

1
bTCPSij ¼ CTCPSij ¼ bTCPSi
1 þ at c ps0 Kpt c ps
  ð6:86Þ
Kpt c ps atcpsi 0
DTCPSij ¼ þ DTCPSi
1 þ at c ps0 Kpt c ps 0 atcpsj

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
at c ps0 at c ps0
DPij ¼ Dastcpsi þ Dxltcpsi
1 þ at c ps0 Kpt c ps 1 þ at c ps0 Kpt c ps
1 ð6:87Þ
þ ðatcpsi DVxyi þ atcpsj DVxyj Þ
1 þ at c ps0 Kpt c ps
¼ dtcpsi Dastcpsi þ ctcpsi Dxltcpsi þ ptcpsi DVxyi þ p0tcpsj DVxyj

Let the transfer function of the TCPS load flow controller be Tltcpsi ðsÞ. It is
assumed that the transfer function of the TCPS stabilizer and feedback signal is
Tltcpsi ðsÞ and the integral of deviation of the active power delivered along the
6.1 Mathematical Model of a Multi-Machine Power System … 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 ¼ Tltcpsi ðsÞ þ Tstcpsi ðsÞ DPij ¼ Ttcpsi ðsÞDPij
s

From Eq. (6.81), it can have

Ttcpsi ðsÞ
D/i ¼ ðatcpsi DVxyi þ atcpsj DVxyj Þ
1 þ at c ps0 Ttcpsi ðsÞ ð6:88Þ
¼ T0tcpsi ðsÞðatcpsi DVxyi þ atcpsj DVxyj Þ

Let the state-space realization of Eq. (6.88) be

X_ TCPSi ¼ ATCPSi XTCPSi þ BTCPSi DVxyi þ BTCPSj DVxyj ð6:89Þ

D/i ¼ CTCPSi XTCPSi þ DTCPSi DVxyi þ DTCPSj DVxyj ð6:90Þ

By substituting Eq. (6.90) into Eq. (6.77), it can be obtained that

DITCSCij ¼ CTCSCij XTCSCi þ DTCSCij DVxyij ð6:91Þ

where

CTCSCij ¼ bTCPSi CTCPSi ;


 
bTCPSi DTCPSi 0
DTCSCij ¼ þ DTCPSi
0 DTCPSj

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 3 2 32 3
I1 Y11 Y12 ... Y1M V1
6 I2 7 6 Y21 Y22 ... Y21 76 V2 7
6 7 6 76 7
6 .. 7¼6 .. .. .. .. 76 .. 7 ð6:92Þ
4 . 5 4 . . . . 54 . 5
IM YM1 YM1 . . . YMM VM
260 6 Multi-machine Power System Installed …

where
     
ixi v g bij
Ii ¼ ; Vi ¼ xi ; Yij ¼ ii ; i; j ¼ 1; 2; . . .; M ð6:93Þ
iyi vyi bji gjj

Let the number of the SVCs, TCSCs, and TCPSs installed in the system be Lsvc ,
Ltcsc and Ltcps , respectively. Denote
 T
DISVC ¼ DITSVC1 DITSVC2 . . . DITSVCLsvc ;
h iT
DITCSC ¼ DITTCSC1j DITTCSC2j . . . DITTCSCLtcsc j ;
h iT
DITCPS ¼ DITTCPS1j DITTCPS2j . . . DITTCPSLtcps j ;
 T ð6:94Þ
DVSVC ¼ DVTSVC1 DVTSVC2 . . . DVTSVCLsvc ;
h iT
DVTCSC ¼ DVTTCSC1j DVTTCSC2j . . . DVTTCSCLtcsc j ;
h iT
DVTCPS ¼ DVTTCPS1j DVTTCPS2j . . . DVTTCPSLtcps j

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 definition 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 ¼ AVSVC XVSVC þ BVSVC DVSVC


ð6:95Þ
DISVC ¼ CSVC XVSVC þ DSVC DVSVC þ bSSVC DaS

X_ LTCSC ¼ ALTCSC XLTCSC þ BTCSC DVTCSC þ bLTCSC DaSTCSC


ð6:96Þ
DITCSC ¼ CTCSC XLTCSC þ DTCSC DVTCSC þ bTCSC DaSTCSC

X_ LTCPS ¼ ALTCPS XLTCPS þ BTCPS DVTCPS þ bLTCPS DaSTCPS


ð6:97Þ
DITCPS ¼ CTCPS XLTCPS þ DTCPS DVTCPS þ bTCPS DaSTCPS
6.1 Mathematical Model of a Multi-Machine Power System … 261

where
2 3 2 3 2 3
XVSVC1 Dxlt csc 1 Dxlt c ps1
6 X 7 6 Dx 7 6 7
6 VSVC2 7 6 lt csc 2 7 6 Dxlt c ps1 7
6 7
XVSVC ¼ 6
6 .. 7; XLTCSC ¼ 6
7 6 .. 7; XLTCPS ¼ 6
7 .. 7;
4 . 5 4 . 5 6 . 7
4 5
XVSVCLsvc Dxlt csc 1 Lt csc xltcpsLtcps
2 3 2 3 2 3
Das1 Dast csc 1 Dastcps1
6 Da 7 6 Da 7 6 7
6 s2 7 6 st csc 2 7 6 Dastcps2 7
6 7
DaS ¼ 6 7
6 .. 7; DaSTCSC ¼ 6
6 .. 7; DaSTCPS ¼ 6
7 .. 7
4 . 5 4 . 5 6 . 7
4 5
DasLsvc Dast csc Ltcsc DastcpsLtcps

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 termi-
nals; (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
2 3 2 32 3
DIxy Ygg Ygs Ygc Ygp Ygo DVxy
6 DISVC 7 6 Ysg Yso 7 6 7
6 7 6 Yss Ysc Ysp 76 DVSVC 7
6 DITCSC 7 ¼ 6 Ycg Ycs Ycc Ycp Yco 7 6 DVTCSC 7 ð6:98Þ
6 7 6 76 7
4 DITCPS 5 4 Ypg Yps Ypc Ypp Ypo 54 DVTCPS 5
0 Yog Yos Yoc Yop Yoo DVO

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 ¼ pSVC DVSVC þ pSVCO DVO


yTCSC ¼ dTCSC DaSTCSC þ cTCSC DXLTCSC þ pTCSC DVTCSC ð6:99Þ
yTCPS ¼ dTCPS DaSTCPS þ cTCPS DXLTCPS þ pTCPS DVTCPS
262 6 Multi-machine Power System Installed …

where
T T
ySVC ¼ ½ DP1j DP1j    DPLsvc j  ; yTCSC ¼ ½ DP1j DP1j    DPLt csc j  ;
 T
yTCPS ¼ DP1j DP1j    DPLcps j ; dTCSC ¼ diagðdt csc i Þ;
cTCSC ¼ diagðct csc i Þ; dTCPS ¼ diagðdtcpsi Þ; cTCSC ¼ diagðctcpsi Þ;
2 0 3
psvc1
6 0 7
2 3 6 p0svc2 7
psvc1 6 7
0 6 .. 7
6 psvc2 7 6 . 7
6 7 6 7
pSVC ¼ 6
6 . 7; pSVCO ¼ 6
7 6 p0svcLsvc 77;
4 . . 5 6 7
0 6 0 7
6 0 7
psvcLsvc 6 7
4 5
..
.

2 3
ptcsc1
6 p0tcsc1 7
6 0 7
6 7
6 ptcsc2 7
6 7
6 p0tcsc2 7
pTCSC ¼6
6
7
7
6 .. 7
6 . 7
6 7
6 7
4 0 ptcscLtcsc 5
p0tcscLtcsc
2 3
ptcps1
6 7
6 p0tcps1 0 7
6 7
6 ptcps2 7
6 7
6 7
6 p0tcsc2 7
pTCPS ¼6 7
6 .. 7
6 . 7
6 7
6 7
6 0 ptcpsLtcps 7
4 5
p0tcpsLtcps

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
6.1 Mathematical Model of a Multi-Machine Power System … 263

 
Cxy X þ Ba Du
0
22 3 2 33
Ygg  Dgxy Ygs Ygc Ygo Ygo
66 7 6 Y 77
66 Ysg Yss  DSVC Ysc Ysp 7 6 so 7 7  
66 7 6 7 7 ½DV
6
¼6 4 Ycg Ycs Ycc  DTCSC Ycp 5 4 Yco 5 7
7 DV
6 7 O
4 Ypg Yps Ypc Ypp  DTCPS Ypo 5
½ og
Y Y os Y oc Yop  Yoo
  
Y0M Y1o ½DV
¼
Y2o Yoo DVO
ð6:100Þ

where
2 3 2 3 2 3
Xg DVxy Cgxy
6X 7 6 DV 7 6 0 7
6 VSVC 7 6 SVC 7 6 CSVC 7
X¼6 7; DV¼6 7; Cxy ¼ 6 7;
4 XLTCSC 5 4 DVTCSC 5 4 CTCSC 5
0
XLTCPS DVTCPS CTCPS
2 3 2 3
Dupss 0
6 Da 7 6 0 7
6 S 7 6 bSSVC 7
Du ¼ 6 7; B a ¼ 6 7
4 DaSTCSC 5 4 bTCSC 5
0
DaSTCPS bTCPS

From Eq. (6.100), it can be obtained that

DVO ¼ Y1
oo Y2o DV
0 ð6:101Þ
Cxy X þ Ba Du ¼ ðYM  Y1o Y1
oo Y2o ÞDV ¼ YM DV

By arranging the state equation (the first 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_ ¼ Axy X þ Bp Du þ Bxy DV ð6:102Þ

where
2 3
Agxy
6 0 7
6 AVSVC 7
Axy ¼ 6 7;
4 ALTCSC 5
0
ALTCPS
2 3 2 3
Bg Bgxy
6 0 7 6 0 7
6 0 7 6 BVSVC 7
Bp ¼ 6 7; Bxy ¼ 6 7
4 bLTCSC 5 4 BTCSC 5
0 0
bLTCPS BTCPS
264 6 Multi-machine Power System Installed …

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_ ¼ AX þ BDu ð6:103Þ

where A ¼ Axy þ Bxy Y1 1


M Cxy ; B ¼ Bp þ Bxy YM Ba . Equation (6.103) is the state
equation of open-loop system without the PSS and thyristor-based FACTS stabi-
lizers 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 ¼ Dx ¼ ½ 0 I 0 Xg ¼ cPSS Xg ð6:104Þ

By using the notation of Eq. (6.100), from Eq. (6.99), the first equation in
Eq. (6.101) and (6.104), it can have

y ¼ C0 X þ D0 Du þ Pxy DV ð6:105Þ

where
2 3 2 3 2 3
yPSS cPSS 0
6y 7 6 0 7 6 0 7
6 SVC 7 0 6 0 7 0 6 0 7
y¼6 7; C ¼ 6 7; D ¼ 6 7;
4 yTCSC 5 4 cTCSC 5 4 dTCSC 5
0 0
yTCPS cTCPS dTCPS
2 3 2 3
0 0
6 0 7 6 p 1 7
6 pSVC 7 6 SVCO Yoo Y2o 7
Pxy ¼6 7þ6 7
4 pTCSC 5 4 0 5
0
pTCPS 0

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 ¼ CX þ DDu ð6:106Þ

where C ¼ C0 þ Pxy Y1 0 1


M Cxy ; D ¼ D þ 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.
6.1 Mathematical Model of a Multi-Machine Power System … 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 ¼ ASVC XSVC þ BSVC DVSVC


ð6:107Þ
DISVC ¼ CSVC XSVC þ DSVC DVSVC þ DSVCO DVO

X_ TCSC ¼ ATCSC XTCSC þ BTCSC DVTCSC


ð6:108Þ
DITCSC ¼ CTCSC XTCSC þ DTCSC DVTCSC

X_ TCPS ¼ ATCPS XTCPS þ BTCPS DVTCPS


ð6:109Þ
DITCPS ¼ CTCPS XTCPS þ DTCPS DVTCPS

where
2 3 2 3 2 3
XSVC1 XTCSC1 XTCPS1
6 XSVC2 7 6 XTCSC2 7 6 XTCPS2 7
6 7 6 7 6 7
XSVC ¼ 6 .. 7; XTCSC ¼ 6 .. 7; XTCPS ¼ 6 .. 7
4 . 5 4 . 5 4 . 5
XSVCLsvc XTCSCLtcsc XTCPSLtcps

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 # 
Cp X Yp Y1po ½DV
¼
0 Y2o Yoo DVO
22 3 2 33
Ygg Dgp Ygs Ygc Ygo Ygo
66 7 6 Y D 77
6 6 Ysg Yss DSVC Ysc Ysp 7 6 so SVCO 7 7 
66 7 6 77 ½DV
¼6
6 4 Ycg Ycs Ycc DTCSC Ycp 5 4 Yco 577 DV
6 7 O
4 Ypg Yps Ypc Ypp DTCPS Ypo 5
½ Yog Yos Yoc Yop  Yoo
ð6:110Þ

where
2 3 2 3
Xg Cgp
6 XSVC 7 0
6 7 6 CSVC 7
X¼4 ;C ¼4 5
XTCSC 5 p CTCSC
0
XTCPS CTCPS
266 6 Multi-machine Power System Installed …

From Eq. (6.110), it can be obtained that


1
DVO ¼ Yoo Y2o DV
0
1
ð6:111Þ
Cp X ¼ ðYp Y1po Yoo Y2o ÞDV ¼ Yp DV

By arranging the state equation (the first equation) in Eqs. (6.107), (6.108),
(6.109), and (6.33) together, it can have

X_ ¼ Ap X þ B0p DV ð6:112Þ

where
2 3
Agp
6 0 7
6 ASVC 7
Ap ¼ 6 7;
4 ATCSC 5
0
ATCPS
2 3
Bgp
6 0 7
6 BSVC 7
B0p ¼ 6 7
4 BTCSC 5
0
BTCPS

From Eqs. (6.111) and (6.112), it can be obtained that

X_ ¼ Aclose X ð6:113Þ

where Aclose ¼ Ap þ 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 Heffron–Phillips Model

Equation (5.43) is the Heffron–Phillips model of the N-machine power system with
the PSSs installed. As this is a linear system satisfying the principle of superim-
position, it only needs to consider the function of one PSS in the discussion. Now, if
6.2 Analysis and Damping Control … 267

K1

-
ω0 I
(sM + D) −1
s
- TPSS
K4 K5
K2

- -
+ Δu pssk
−1 −1 +
(K 3 + sTd0 ) (I + sTA ) K A Fk
-

K6

Fig. 6.11 Heffron–Phillips model of a multi-machine power system installed with a PSS

only the kth PSS is considered, the Heffron–Phillips model of the power system can
be obtained from Eq. (5.43) as

sDd ¼ xo IDx
sDx ¼ M1 ðK1 Dd  K2 DE0q  DDxÞ
ð6:114Þ
sDE0q ¼ T01 0 0
d0 ðK3 DEq  K4 Dd þ DEfd Þ

sDE0fd ¼ T1 0 1 0
A DEfd þ TA KA ðK5 Dd  K6 DEq þ Fk Dupssk Þ

where Fi ¼ ½ 0 ... 0 1 0 . . . 0 T and Dupssk is the output stabilizing


"
the kth element
signal from the kth PSS. From Eq. (6.114), the Heffron–Phillips 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 infinite-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
¼ M1 K2 ½ðI þ sTA ÞðK3 þ sTd0 Þ þ KA K6 1 KA Fk ¼ FPSS ðsÞ ð6:115Þ
Dupssk

Let the transfer function and feedback signal of the PSS be Tpssk ðsÞ and Dyk ,
respectively, that is

Dupssk ¼ Tpssk ðsÞDyk ð6:116Þ

The feedback signal Dyk can always be reconstructed by using Dxj ; j ¼


1; 2; . . .; N (see Sect. 6.2.1.2), that is
268 6 Multi-machine Power System Installed …

Dyk ¼ cj ðsÞDxj ; j ¼ 1; 2; . . .; N ð6:117Þ

If the low-frequency power oscillation of interests is associated with the ith


i ¼ n þ jxi , the electric torque contribution
oscillation mode of the power system, k i
of the PSS at the complex angular oscillation frequency k i ¼ n þ jxi can be
i
obtained from Eqs. (6.115) to (6.117) as
2 3
 i ÞT
pss1 ðk  pssk ðk i Þc ðk

F 1 i ÞDx1
6     7
6  7
DTPSS ¼F i )T
 PSS ðk i ÞDy ¼ 6 Fpss2 ðki ÞTpssk ðki Þc2 ðki ÞDx2 7
 pssk ðk ð6:118Þ
k 6 .. 7
4 . 5
 i ÞT
pssN ðk
F  pssk ðki Þc ðki ÞDxN
N

where F i Þ; j ¼ 1; 2; . . .N is the jth element of F


pssj ðk i Þ. Hence, the damping
PSS ðk
torque contribution from the PSS to the electromechanical oscillation loop of each
generator is

 i ÞT
~pssj ðk
DTdampj ¼ Ddampj Dxj ¼ Re½F i Þc ðk
 pssk ðk 
j i ÞDxj ð6:119Þ

From the unified Heffron–Phillips 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
¼ FFACTS ðs)
Dufactss
¼ M1 KP  M1 K2 ½ðI þ sTA ÞðK3 þ sTd0 Þ þ KA K6 1 ½ðI þ sTA ÞKq þ KA KV 
ð6:120Þ

Similarly, let the transfer function and feedback signal of the thyristor-based
FACTS stabilizer be Tfacts ðsÞ and Dyk , respectively, that is

Dufactss ¼ Tfacts ðs)Dyk ð6:121Þ

The electric torque contribution of the thyristor-based FACTS stabilizer at the


complex angular oscillation frequency k i ¼ n þ jxi can be obtained from
i
Eqs. (6.117), (6.120), and (6.121) as
2 3
 i ÞT
facts1 ðk
F ~ facts ðki Þ i ÞDx1
c1 ðk
6F  i ÞT
facts2 ðk  facts ðki Þ  ÞDx2 7
c2 ðk
DTFACTS ¼F i ÞT
 FACTS ðk i ÞDy ¼ 6
 facts ðk 6 .
i 7
7 ð6:122Þ
k
4 .. 5
 i ÞT
factsN ðk
F  facts ðki Þ
c ðki ÞDxN
N
6.2 Analysis and Damping Control … 269

where F i Þ; j ¼ 1; 2; . . .; N is the jth element of F


factsj ðk i Þ. Hence, the
 FACTS ðk
damping torque contribution from the thyristor-based FACTS stabilizer to the
electromechanical oscillation loop of each generator is

 i ÞT
factsj ðk
DTdampj ¼ Ddampj Dxj ¼ Re½F i Þc ðk
 facts ðk 
j i ÞDxj ð6:123Þ

6.2.1.2 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_ ¼ AX þ BDu ð6:124Þ

where
2 3 2 3 2 3
Xg Dupss Dupss1
6X 7 6 Da 7 6 Dupss2 7
6 VSVC 7 6 7 6 7
7; Dupss ¼ 6 7;
S
X¼6 7; Du ¼ 6 6 . 7
4 XLTCSC 5 4 DaSTCSC 5 4 .. 5
XLTCPS DaSTCPS DupssN
2 3 2 3 2 3
Das1 Dastcsc1 Dastcps1
6 Da 7 6 Da 7 6 7
6 s2 7 6 stcsc2 7 6 Dastcps2 7
6 7 6 7 6 7
DaS ¼ 6 . 7; DaSTCSC ¼ 6 .. 7; DaSTCPS ¼ 6 .. 7
4 .. 5 4 . 5 6
4 . 7
5
DasLsvc DastcscLtcsc DastcpsLtcps

Equation (6.124) can be rearranged into the following form:


2 3 2 32 Dd 3 2 3
Dd_ x0 I 0
6 Dx_ 7 4
0 0
6 7 X 6 7
4 5 ¼ A21 A22 A23 54 Dx 5 þ 4 B2k 5Duk ð6:125Þ
DZ_ A31 A32 A33 DZ k
B3k

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 ¼ Dupssj ; j ¼ 1; 2; . . .; N; Duk ¼ Dasj ; j ¼ 1; 2; . . .; Lsvc ;


Duk ¼ Dastcscj ; j = 1,2,. . .; Ltcsc ; Duk ¼ Dastcpsj ; j ¼ 1; 2; . . .; Ltcps
270 6 Multi-machine Power System Installed …

A 21

A 22

Δ Δ
1 ω0 I
+
s s
Δu k
B 2k B3k A 32 A 31

A 23

1
Δz +
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
!
1
X
DZ ¼ ðsI  A33 Þ A31 Dd þ A32 Dx þ B3k Duk ð6:126Þ
k

Substituting the above equation into Eq. (6.125), it can have

Dd_ ¼ x0 Dx
Dx_ ¼ ½A21 þ A23 ðsI  A33 Þ1 A31 Dd þ ½A22 þ A23 ðsI  A33 Þ1 A32 Dx
X
þ ½B2k þ A23 ðsI  A33 Þ1 B3k Duk ð6:127Þ
k
X
¼ A21 ðsÞDd þ A22 ðsÞDx þ Bk ðs)Duk
k

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 ðsÞ and
Dyk , respectively, that is
6.2 Analysis and Damping Control … 271

A 21 (s)

A 22 (s)

Δ Δ
1 ω0 I
+
s s

Δu k
Σ B k (s)

Fig. 6.13 Compact form of general linearized model of an N-machine power system

Duk ¼ Tstk ðsÞDyk ð6:128Þ

From Fig. 6.13, Eqs. (6.117), and (6.128), the electric torque contribution from
the kth stabilizer can be obtained as
2 3
 i ÞT
 k1 ðk i Þc ðk
 stk ðk 
B 1 i ÞDx1
6 7
6      7
 i ÞT
 k ðk
DTST ¼ B i ÞDy ¼ 6 Bk2 ðki ÞTstk ðki Þc2 ðki ÞDx2 7
 stk ðk k 6 .. 7
4 . 5
    
BkN ðki ÞTstk ðki Þc ðki ÞDxNN

i Þ; j ¼ 1; 2; . . .; N is the jth element of Bk ðk


where Bkj ðk i Þ. Hence, the damping
torque contribution from the kth stabilizer to the electromechanical oscillation loop
of each generator is (for j ¼ 1; 2; . . .; N)

i ÞT
 kj ðk
DTdampj ¼ Ddampj Dxj ¼ Re½B i Þc ðk
 stk ðk 
j i ÞDxj ð6:129Þ

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 ¼ 1; 2; . . .; N)
2 3 2 32 Dd 3 2 0 3
Dd_ j 0 x0 0 j
6 _ 7 4 6 7 6 7
4 D x j5 ¼ k j dj Aj 54 Dxj 5 þ 4 bjk 5Duk ð6:130Þ
DZ_ j Aj1 Aj2 Aj3 DZj Bjk

From Eqs. (6.117), (6.128), and (6.130), it can be obtained that


hx i
DZj ¼ ðsI  Aj3 Þ1
0
Aj1 þ Aj2 þ Bjk Tstk ðsÞcj ðsÞ Dxj ð6:131Þ
s

The output equation of the system about the feedback signal of the stabilizer can
be written as
272 6 Multi-machine Power System Installed …

2 3
Ddj
 6 7
Dyk ¼ c1kj c2kj C3kj 4 Dxj 5 þ djk Duk ð6:132Þ
DZj

By using Eqs. (6.117), (6.128), and (6.131), Eq. (6.131) becomes


x
1 x0
0
Dyk ¼ c1kj þ c2kj Dxj þ C3kj sI  Aj3 Aj1 þ Aj2 Dxj
s s ð6:133Þ
 ½C3kj ðsI  Aj3 Þ1 Bjk þ djk Tstk ðsÞcj ðsÞDxj

From Eqs. (6.117) and (6.133), it can have

Dyk x0 x
c1kj þ c2kj þ C3kj ðsI  Aj3 Þ1
0
cj ðsÞ ¼ ¼ Aj1 þ Aj2
Dxj s s ð6:134Þ
 ½C3kj ðsI  Aj3 Þ1 Bjk þ djk Tstk ðsÞcj ðsÞ

Hence, finally, it can have

ðx0 c1kj þ c2kj Þ þ C3kj ðsI  Aj3 Þ1 ðxs0 Aj1 þ Aj2 Þ
cj ðsÞ ¼ s ð6:135Þ
½C3kj ðsI  Aj3 Þ1 Bjk þ djk Tstk ðsÞ

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).
Denote the right and left eigenvectors of the state matrix with respect to the mode k i
to be (see Sect. 2.2.1 and 5.2.1.1)
2 3
v1i  
vi ¼ 4 v2i 5; w Ti ¼ w 1i  2i
w  T3i
w ð6:136Þ
v3i
6.2 Analysis and Damping Control … 273

The following participation factor (see Eq. (5.60)

jw 2i v2i j
pji ¼ T ð6:137Þ
 i vi
w

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-defined participation factor, the jth generator is chosen to be
the best installing location of a PSS to suppress the power system oscillation
associated with the oscillation mode ki . The general procedure of the selection is to
arrange the state matrix of the system according to Eq. (6.125), that is
2 3
0 x0 I 0
A ¼ 4 A21 A22 A23 5 ð6:138Þ
A31 A32 A33

Denote the right and left eigenvectors of the above state matrix with respect to
i to be
the mode k
2 3
v1i  T 
vi ¼ 4 v2i 5; w
 Ti ¼ w
 1i  T2i
w  T3i
w ð6:139Þ
v3i

The participation factors of all generators can be calculated as



vj2i uj2i

pji ¼ T ðfor j ¼ 1; 2; . . .; NÞ ð6:140Þ
 v2i
w2i

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 artificial damping in the electromechanical
oscillation loop of each generator in the power system, Ddj Dxj (for j ¼ 1; 2; . . .; N).
If the maximum improvement, Dk i , is
i , for the oscillation mode of interests, k
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 coefficient
of the added extra damping torque in the electromechanical oscillation mode, Ddj .
That is
274 6 Multi-machine Power System Installed …

 
ij ¼ Dki ¼ @ ki
S ð6:141Þ
Ddj @dj

From Eqs. (6.130) and (6.136), it can have


2 32 3
   0 0 0 v1i
ij ¼ @ ki ¼ w
S  Ti
@A
vi ¼ w 1i  2i
w  T3i 4 0
w 1 0 54 v2i 5 ¼ w
 2i v2i ð6:142Þ
@dj @dj v3i
0 0 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.
Let the oscillation mode be k i ¼ n þ jxi . The sensitivity index used is
i

Dni
Sijn ¼ ð6:143Þ
Ddj

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 defined in Eqs. (6.141) or
(6.143), the addition of damping torque, Ddampj Dxj (for j ¼ 1; 2; . . .; N), on each
generator affects the oscillation mode of interests. Hence, the total improvement of
the oscillation mode of interests ki by the stabilizer can be obtained as

X
N X
N
i ¼
Dk ij ; or Dni ¼
Ddampj S Ddampj Sijn ð6:144Þ
j¼1 j¼1
6.2 Analysis and Damping Control … 275

Generator 1

Ddamp1 Si1

Generator 2
Ddamp2 Si2
The
oscillation
A
mode of
stabilizer
interests

DdampN SiN
Generator N

Damping torque contribution Damping torque contribution is


to each generator 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 Heffron–Phillips 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:
X
N
DTA ¼ F i Þc ðk
pssj ðk  
j i ÞSij ðfor a PSS using HeffronPhillips modelÞ ð6:145Þ
j¼1

X
N
DTA ¼ F i Þc ðk
fsctsj ðk  
j i ÞSij ð6:146Þ
j¼1
276 6 Multi-machine Power System Installed …

(for a thyristor-based FACTS stabilizer using Heffron–Phillips model)

X
N
DTA ¼ B i Þc ðk
 kj ðk 
j i ÞSij ð6:147Þ
j¼1

(for a PSS or thyristor-based FACTS stabilizer using general linearized model)


where cj ðsÞ ¼ ðxs0 c1kj þ c2kj Þ þ C3kj ðsI  Aj3 Þ1 ðxs0 Aj1 þ Aj2 Þ which is given by
Eq. (6.135).
In the above DTA indices, often S ij is replaced by Sijn so as to focus the
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 3 2 32 Dd 3 2 3
Dd_ 0 x0 I 0 0
6 7 76 7 6
6 Dx_ 7 ¼ 6 A22 A23 56 7 7
4 5 4 A21 4 Dx 5 þ 4 B2k 5Duk
DZ_ A31 A32 A33 DZ B3k
ð6:148Þ
2 3
Dd
 
T 6 7
Dyk ¼ CT1k CT2k C3k 4 Dx 5
DZ

where Dyk is the feedback signal of the stabilizer. If the right and left eigenvectors
i , are vi and w
of the state matrix with respect to the oscillation mode of interests, k  Ti ,
respectively, the controllability, observability, and residue index of the stabilizer are
(see Sect. 2.2.1)
2 3
0  
  Ti 4 B2k 5; cik ¼ CT1k
bik ¼ w CT2k  ik ¼ bikcik
CT3k vi ; R ð6:149Þ
B3k

Figure 6.15 shows that cik measures how much the oscillation mode k i is
observed in Dyk and bik measures how much the stabilizing signal Duk affects k i .

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.
6.2 Analysis and Damping Control … 277

Fig. 6.15 Configuration of


1
the system in modal bik cik
decomposition s − λi

Channels associated with other


eigenvalues of the system
+

Δu k Δy 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.
From Eq. (2.60), it can have
2 3
2 3 z1 ð0Þek1 t
Dd 6 z2 ð0Þek2 t 7
4 Dx 5 ¼ V6
6 ..
7
7 ð6:150Þ
4 . 5
DZ
zn ð0Þekn t

Hence,

X
N
v2ij zi ð0Þ
Dxj ðsÞ ¼ i ð6:151Þ
i¼1 sk

v2ij is the jth element of vi2 in Eq. (6.139). While


where 
2 3
2 3 z1 ð0Þek1 t
Dd  6 k2 t 7
   6 z2 ð0Þe 7
Dyk ¼ CT1k CT2k CT3k 4 Dx 5 ¼ CT1k CT2k CT3k V6 .. 7
4 . 5
DZ
zn ð0Þekn t

Hence,

X
N
cik zi ð0Þ
Dyk ðsÞ ¼ i ð6:152Þ
i¼1 sk
278 6 Multi-machine Power System Installed …

From Eqs. (6.150) and (6.152), it can have


PN c ik zi ð0Þ i Þ PN c ik zið0Þ
Dyk ðsÞ i¼1 sk i ðs  k i¼1 ski
¼ ¼ ð6:153Þ
Dxj ðsÞ PN v 2ij zi ð0Þ ðs  k i Þ PN v 2ij zi ð0Þ
i¼1 ski i¼1 ski

i and using Eq. (6.117), it can be obtained


In the above equation, by letting s ¼ k
that

i Þ cik zi ð0Þ
Dyk ðk cik
i Þ ¼
cj ðk
i Þ ¼ v2ij zi ð0Þ ¼ v2ij
Dxj ðk
ð6:154Þ

On the other hand, from Eqs. (6.139) and (6.148), it can have
2 3
  0 x0 I 0  T 
 T1i
w  T2i
w  T3i 4 A21
w A22 i w
A23 5 ¼ k  1i  T2i
w  T3i
w ð6:155Þ
A31 A32 A33

The above equation gives

w  T3i A33 ¼ w
 T2i A23 þ w i
 T3i k ð6:156Þ

Solution of Eq. (6.156) is

 T3i ¼ w
w i I  A33 Þ1
 T2i A23 ðk ð6:157Þ

Hence, from Eqs. (6.148), (6.149), and (6.157), it can have


2
3 3 2
0 0
 6 7  T 6 7
 Ti 4 B2k 5 ¼ w
bik ¼ w  1i  T2i w
w  T3i 4 B2k 5 ¼ w  T2i B2k þ w
 T3i B3k
B3k B3k ð6:158Þ
X
¼w T i I  A33 Þ B3k  ¼ w
 2i ½B2k þ A23 ðk 1 i Þ ¼
 T2i Bk ðk i Þ
 2ij Bkj ðk
w
j

From Eqs. (6.139), (6.142), and (6.148), it can have


2 3
0 x0 I 0
@ 4 A21 A22 þ diagðdj Þ A23 5
i
@k A31 A32 A33
Sij ¼ i
¼w T
vi ¼ w
 2ij v2ij ð6:159Þ
@dj @dj

Hence, from Eqs. (6.154), (6.158), and (6.159), the residue index can be
obtained as
6.2 Analysis and Damping Control … 279

X X
 ik ¼ 
bikcik ¼ i Þc ðk
 kj ðk
 2ij B  ij B i Þc ðk
 kj ðk 
R w j i Þv2ij ¼ S j iÞ ð6:160Þ
j j

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 sta-
bilizer. 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 UðuÞ and that of power system
operating conditions XðlÞ. The effectiveness of the stabilizer is the function of φ
and μ, C(φ, μ). 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

Max½Cðu; l0 Þ; l0 2 XðlÞ ð6:161Þ


u

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.

Max½min Cðu; lÞ ð6:162Þ


u l

Min½max Cðu; lÞ  min Cðu; lÞ ð6:163Þ


u l l

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 6 Multi-machine Power System Installed …

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 significantly, 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 effec-
tiveness 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:
2 3 2 32 Dd 3 2 3
Dd_ j 0 x0 j0 0
6 7 76 7 6
6 Dx_ j 7 ¼ 6 Aj 56 7 7
4 5 4 kj dj 4 Dxj 5 þ 4 bjk 5Duk
DZ_ j Aj1 Aj2 Aj3 DZj Bjk
2 3 ð6:164Þ
Ddj
 6 7
Dyk ¼ c1kj c2kj C3kj 6 7
4 Dxj 5
DZj

Let the right and left eigenvectors of the state matrix, respectively, be
2 3
v1i  
vi ¼ 4 v2i 5; w Ti ¼ w 1i  2i
w  T3i
w ð6:165Þ
v3i
6.2 Analysis and Damping Control … 281

According to the definition of the right and left eigenvectors, it can have
2 3
0 x0 0
 6 7  
 1i
w  2i
w  T3i 4 kj
w dj Aj 5 ¼ k  1i
i w  2i
w  T3i
w
Aj1 Aj2 Aj3
2 32 3 2 3 ð6:166Þ
0 x0 0 v1i v1i
6 76 7  6 7
4 kj dj Aj 54 v2i 5 ¼ ki 4 v2i 5
Aj1 Aj2 Aj3 v3i v3i

From Eq. (6.166), it can be obtained that

 Ti3 ¼ Aj ðk
w i I  Aj3 Þ1 w
 i2
x0
vi1 ¼  vi2
ki ð6:167Þ
 
i I  Aj3 Þ1 Aj1 x0 þ Aj2 vi2
vi3 ¼ ðk i
k

From Eq. (6.167), controllability and observability index can be obtained,


respectively, as
2 3
0
 6 7
bik ¼ w 1i  2i
w  T3i 4 bjk 5
w
Bjk
i I  Aj3 Þ1 Bjk w
¼ ½bjk þ Aj ðk  2i ¼ K  i Þw
 bi ðk  i2
2 3
v1i
 6 7
cik ¼ c1kj c2kj C3kj 4 v2i 5
v3i
 
x0  1 x0 i Þv2i
 ci ðk
¼  c1kj þ c2kj þ C3kj ðki I  Aj3 Þ ðAj1  þ Aj2 Þ v2i ¼ K
ki ki
ð6:168Þ

Hence, the equivalent residue index is

 ik ¼ bikcik ¼ K
R  i ÞK
 bi ðk i Þvi2 w
 ci ðk  i2 ð6:169Þ

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 com-
parison 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
jR  ikB j, then A is considered to be better than B to be the installing location
282 6 Multi-machine Power System Installed …

R 
or the feedback signal of the stabilizer. Therefore, it is the ratio,  , not the
ikA

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 ¼ vi2B ; w
 i2A ¼ w
 i2B ð6:170Þ

Therefore, from Eqs. (6.169) and (6.170), it can be obtained that



 ikA j K
jR  i ÞK
 biA ðk i Þvi2A w
 ciA ðk  i2A

 ikB j ¼ K
jR  i ÞK
 biB ðk i Þvi2B w
 ciB ðk  i2B
ð6:171Þ
K  i ÞK
 biA ðk i Þ jvi2A w
 ciA ðk  i2A j K  i ÞK
 biA ðk  i Þ
 ciA ðk
¼ ¼
 i ÞK
 biB ðk
K i Þ jvi2B w
 ciB ðk  i2B j K  i ÞK
 biB ðk  i Þ
 ciB ðk

Equation (6.171) shows that jK  bi ðki ÞK