TRANSIENT STABILITY ANALYSIS OF MULTIMACHINE POWER SYSTEMS

BY CATASTROPHE THEORY

by

A l l MOHAMED MIHIRIG
B . S c , U n i v e r s i t y of A l Fateh, 1978
M.A.Sc, U n i v e r s i t y of B r i t i s h Columbia, 1984

A THESIS SUBMITTED IN PARTIAL FULFILMENT OF

THE REQUIREMENTS FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

in

THE FACULTY OF GRADUATE STUDIES

DEPARTMENT OF ELECTRICAL ENGINEERING

We accept t h i s t h e s i s as conforming

to t h e r e q u i r e d standard

THE UNIVERSITY OF BRITISH COLUMBIA

December 1987

Ali Mohamed M i h i r i g , 1987

 In presenting this thesis in partial fulfilment of the requirements for an advanced

degree at the University of British Columbia, I agree that the Library shall make it

freely available for reference and study. I further agree that permission for extensive

copying of this thesis for scholarly purposes may be granted by the head of my

department or .by his or her representatives. It is understood that copying or

publication of this thesis for financial gain shall not be allowed without my written

permission.

Department

The University of British Columbia

1956 Main Mall
Vancouver, Canada
V6T 1Y3

Date /4d\rcl\ <S y

DE-6G/81)
 ABSTRACT

Transient stability analysis i s an Important part of power system

planning and operation. For large power systems, such analysis i s very

demanding i n computation time. On-line transient stability assessment will

be n e c e s s a r y f o r secure and r e l i a b l e o p e r a t i o n of power systems i n the near

f u t u r e because systems are o p e r a t e d c l o s e t o t h e i r maximum limits.

In the l a s t two decades, a v a s t amount of r e s e a r c h work has been done

in the area of f a s t transient stability assessment by d i r e c t methods. The

major d i f f i c u l t i e s a s s o c i a t e d w i t h d i r e c t methods are the l i m i t a t i o n s i n the

power system model, determination of transient stability regions and

adaptation to changes i n operating conditions. In t h i s thesis catastrophe

t h e o r y i s used to determine the t r a n s i e n t stability regions. Taylor series

expansion i s used to f i n d the energy balance e q u a t i o n i n terms of clearing

time and system transient parameters. The energy function i s then put i n

the form of a catastrophe manifold from which the bifurcation set is

extracted. The b i f u r c a t i o n s e t r e p r e s e n t s the t r a n s i e n t s t a b i l i t y region i n

terms of the power system transient parameters bounded by the transient

stability limits. The transient stability r e g i o n s determined are v a l i d for

any changes in loading conditions and fault location. The transient

stability problem i s d e a l t w i t h i n the two dimensions of t r a n s i e n t stability

limits and c r i t i c a l c l e a r i n g times. Transient s t a b i l i t y limits are g i v e n by

the bifurcation s e t and the c r i t i c a l clearing times are c a l c u l a t e d from the

catastrophe manifold equation. The method a c h i e v e s a breakthrough i n the

modelling problem because the e f f e c t s of e x c i t e r response, f l u x decay and

- i i -
systems damping can a l l be included i n the transient stability analysis.

Numerical examples of one-machine infinite-bus and multi-machine power

systems show v e r y good agreement with the time solution i n the p r a c t i c a l

range of f i r s t swing s t a b i l i t y analysis. The method p r e s e n t e d f u l f i l l s a l l

requirements for on-line assessment of transient stability of power

systems.

- i i i -
 TABLE OF CONTENTS

Page

ABSTRACT i i

TABLE OF CONTENTS iv

LIST OF TABLES v i i

LIST OF FIGURES viii

ACKNOWLEDGEMENT xi

CHAPTER 1 INTRODUCTION 1

CHAPTER 2 REVIEW OF THE STABILITY PROBLEM AND ITS SOLUTION 5

2.1 Introduction 5

2.2 The S t a b i l i t y Problem 5

2.3 B a s i c Power System S t a b i l i t y Model 7

2.3.1 Synchronous Generator 8

2.3.2 E x c i t e r and Governor C o n t r o l Systems 10

2.3.3 T r a n s m i s s i o n System and Loads 11

2.4 S o l u t i o n of the T r a n s i e n t S t a b i l i t y Problem 12

2.4.1 Numerical I n t e g r a t i o n 12

2.4.2 Lyapunov's D i r e c t Method 15

2.4.3 Pattern Recognition Method 18

2.5 D i s c u s s i o n of E x i s t i n g Methods 22

CHAPTER 3 APPLICATION OF CATASTROPHE THEORY TO TRANSIENT

STABILITY ANALYSIS 26

3.1 Introduction ' 26

- iv -
 Page

3.2 The C a t a s t r o p h e Theory 27

3.2.1 C a t a s t r o p h e Theory and B i f u r c a t i o n

Analysis 28

3.3 A p p l i c a t i o n to the Steady S t a t e Stability

Problem 29

3.3.1 C y l i n d r i c a l - R o t o r I n f i n i t e Bus Power

System 31

3.3.2 Numerical R e s u l t s 38

3.4 A p p l i c a t i o n to the T r a n s i e n t S t a b i l i t y Problem 38

3.4.1 S i n g l e Machine I n f i n i t e - b u s Power System. 41

3.4.2 Results 48

3.4.3 E q u i v a l e n t Two Machine Power System 52

3.5 Summary 54

CHAPTER 4 TRANSIENT STABILITY REGIONS OF MULTI-MACHINE

POWER SYSTEMS 55

4.1 Introduction 55

4.2 Dynamic E q u i v a l e n t of the C r i t i c a l Machines 56

4.2.1 The T r a n s i e n t S t a b i l i t y Regions 61

4.3 T a y l o r E x p a n s i o n of the A c c e l e r a t i n g Power 65

4.3.1 T r a n s i e n t S t a b i l i t y Regions U s i n g T a y l o r

Expansion 70

4.4 I d e n t i f i c a t i o n of the C r i t i c a l Machines 72

4.5 N u m e r i c a l Examples 74

4.5.1 The Three-Machine System 75

4.5.2 CIGRE 7-Machine T e s t System 84

- -v
 Page

4.6 D i s c u s s i o n of R e s u l t s 84

CHAPTER 5 INCLUSION OF DAMPING, FLUX DECAY AND EXCITATION

RESPONSE 90

5.1 L i m i t a t i o n of the C l a s s i c a l Model 90

5.2 Damping 91

5.3 E x c i t a t i o n Response and F l u x Decay 95

CHAPTER 6 CONCLUSION 107

REFERENCES 112

APPENDIX A 117

APPENDIX B 122

APPENDIX C 126

APPENDIX D 130

APPENDIX E 131

- vi -
 LIST OF TABLES

Page

TABLE 3.1 The 7-elementary c a t a s t r o p h e s 30

TABLE 4.1 Network parameters of the three-machine power system. 75

TABLE 4.2 Generator data and i n i t i a l c o n d i t i o n s of the

three-machine power system 77

TABLE 4.3 Critical clearing time by the time s o l u t i o n and

the proposed methods f o r the 3-machine power

system 80

TABLE 4.4 Data of the CIGRE 7-machine system 85

TABLE 4.5 Critical clearing time by the time s o l u t i o n and

the proposed methods f o r the 7-machine power

system 86

- vii -
 LIST OF FIGURES

Page

F i g u r e 2.1 P a t t e r n r e c o g n i t i o n c l a s s i f i c a t i o n procedure 19

F i u g r e 2.2 One-machine i n i f i n i t e - b u s power system 24

F i g u r e 2-3 C h a r a c t e r i s t i c s of KE and TM w i t h load v a r i a t i o n 24

F i g u r e 2.4 The d e c i s i o n f u n c t i o n s f o r t h r e e d i f f e r e n t fault

locations 24

F i g u r e 3.1 One-machine i n f i n i t e - b u s power system 32

F i g u r e 3.2 The power angle c u r v e 32

F i g u r e 3.3 The b i f u r c a t i o n s e t of the cusp c a t a s t r o p h e which

represents the steady s t a t e s t a b i l i t y criteria 37

F i g u r e 3.4 The energy f u n c t i o n s t a b i l i t y criteria 39

F i g u r e 3.5 One machine i n f i n i t e - b u s power system 47

F i g u r e 3.6 The t r a n s i e n t s t a b i l i t y l i m i t s g i v e n by the

swallowtail catstrophe 47

F i g u r e 3.7 Transient stability region f o r a l l possible fault

l o c a t i o n s and l o a d i n g c o n d i t i o n s 49

F i g u r e 3.8 The t r a n s i e n t s t a b i l i t y r e g i o n i n terms of s t a b i l i t y

l i m i t s and c r i t i c a l c l e a r i n g times 50

F i g u r e 3.9 The e q u i v a l e n t two machine power system 53

F i g u r e 4.1 The three-machine power system (WSCC) 76

F i g u r e 4.2 The accuracy of T a y l o r s e r i e s d u r i n g the f a u l t on

p e r i o d u s i n g only the second order term 78

- viii -
 Page

F i g u r e A.3 The a c c e l e r a t i o n power f o r d i f f e r e n t fault locations. 79

F i g u r e 4.4 The t r a n s i e n t s t a b i l i t y r e g i o n of the three-machine

power system u s i n g the dynamic e q u i v a l e n t method 81

F i g u r e 4.5 The t r a n s i e n t s t a b i l i t y r e g i o n f o r the three-machine

system by the R a y l o r s e r i e s method 82

F i g u r e 4.6 The CIGRE t e s t system 83

F i g u r e 4.7 The t r a n s i e n t s t a b i l i t y r e g i o n f o r the CIGRE test

system u s i n g the dynamic e q u i v a l e n t method 87

F i g u r e 4.8 The t r a n s i e n t s t a b i l i t y r e g i o n f o r the CIGRE test

system u s i n g the T a y l o r s e r i e s method 88

F i g u r e 5.1 Effect of damping i n c l u s i o n on the a c c e l e r a t i n g

power d u r i n g f a u l t - o n p e r i o d 96

F i g u r e 5.2 Effect of damping i n c l u s i o n on the a c c e l e r a t i n g

power d u r i n g f a u l t - o n p e r i o d 96

F i g u r e 5.3 The phasor diagram of the g e n e r a t o r internal

q u a n t i t i e s f o r the t r a n s i e n t s t a t e 98

F i g u r e 5.4 The t y p i c a l e x c i t a t i o n v o l t a g e - t i m e response 100

F i g u r e 5.5 The e f f e c t of f l u x decay and e x c i t a t i o n response on

the a c c e l e r a t i n g power d u r i n g the t r a n s i e n t p e r i o d . . . 106

- ix -
 Page

Figure A . l Cross s e c t i o n of the s w a l l o w t a i l b i f u r c a t i o n s e t f o r

U > 0 120

F i g u r e A.2 Cross s e c t i o n of the s w a l l o w t a i l b i f u r c a t i o n s e t f o r

U < 0 120

F i g u r e A.3 The 3-D b i f u r c a t i o n s e t of the swallowtail

catastrophe 120

Figure B.l The cusp m a n i f o l d and i t s b i f u r c a t i o n set 123

F i g u r e B.2 The cusp p o t e n t i a l f u n c t i o n V(X) at d i f f e r e n t v a l u e s

of c o n t r o l v a r i a b l e s 125

x -
 ACKNOWLEDGEMENT

I would l i k e to thank Dr. M.D. Wvong f o r h i s c o n t i n u a l encouragement

and i n v a l u a b l e guidance throughout the course of t h i s research.

I would like to express my sincere appreciation and indebteness to

Dr. H.W. Dommel f o r h i s encouragement and f i n a n c i a l assistance.

I am a l s o indebted to the B.C. Hydro and Power A u t h o r i t y for the

financial support and t o the E l e c t r i c a l Engineering Department f o r the

computing support and the t e a c h i n g a s s i s t a n t s h i p s provided.

I wish to thank my w i f e , Najwa, f o r her encouragement and p a t i e n c e

throughout my graduate program.

Thanks are extended t o G a i l Schmidt f o r her p l e a s a n t a t t i t u d e and

skilled t y p i n g of the m a n u s c r i p t .

- xi-
 1

CHAPTER 1

INTRODUCTION

The stability of a power system implies that a l l i t s generators

remain i n synchronism through normal and abnormal operating conditions.

Transient stability arises when a large disturbance such as a loss of

generation, load or t r a n s m i s s i o n l i n e s takes p l a c e i n the power system. The

question of whether the power system w i l l settle t o a new stable operating

s t a t e or not i s known as the t r a n s i e n t s t a b i l i t y problem.

The ever i n c r e a s i n g demand f o r e l e c t r i c a l energy r e q u i r e s ever l a r g e r

interconnected power systems and a maximum g e n e r a t i o n . This raises great

concern about the s e c u r i t y of power systems when s u b j e c t e d to large d i s t u r -

bances. Transient stability, therefore, becomes an i n c r e a s i n g l y important

consideration i n system p l a n n i n g and o p e r a t i o n . Extensive stability studies

are needed i n order t o ensure system s e c u r i t y b e f o r e a planning or o p e r a t i n g

d e c i s i o n i s made. Each c o n t i n g e n c y f o r each d i s t u r b a n c e considered requires

a l a r g e number of s t a b i l i t y studies to determine the c r i t i c a l c l e a r i n g time

or system s t a b i l i t y limits.

A typical transient stability study consists of o b t a i n i n g the time

solution to the power system d i f f e r e n t i a l and a l g e b r a i c equations s t a r t i n g

with the system c o n d i t i o n s p r i o r to the t r a n s i e n t . The power system equa-

tions should include a l l significant parameters that influence stability

such as generator c o n t r o l s , s t a b i l i t y c o n t r o l s and p r o t e c t i v e d e v i c e s . The

d e s i r e d o b j e c t i v e s of a t r a n s i e n t s t a b i l i t y analysis are:
 2

i. The s t a b i l i t y of the power system. I s i t s t a b l e or n o t , t o what degree

is i t s t a b l e , and how f a r i s i t from the s t a b i l i t y limits?

ii. Time responses of generator v a r i a b l e s , bus v o l t a g e s , c u r r e n t s , and

a c t i v e and r e a c t i v e power.

iii. System q u a n t i t i e s t h a t a f f e c t the performance of p r o t e c t i v e d e v i c e s .

The above objectives are key i s s u e s i n power system design, p l a n n i n g and

o p e r a t i o n t o ensure system s t a b i l i t y f o r different prescribed disturbances.

The time s o l u t i o n method of s t a b i l i t y a n a l y s i s , although i t i s very

reliable, accurate, and s u i t a b l e f o r d i f f e r e n t modelling orders, has the

following disadvantages;

1. The method i n v o l v e s n u m e r i c a l i n t e g r a t i o n of a l a r g e number of d i f f e r -

ential equations f o r each disturbance considered. A large number of

repetitive simulations i s required f o r each case t o determine either

the stability limits or the c r i t i c a l clearing time. T h i s procedure I s

very time consuming i n the system p l a n n i n g stage where a l a r g e number

of cases need t o be c o n s i d e r e d .

2. I n system o p e r a t i o n s , t h e r e a r e s i t u a t i o n s where f a s t s o l u t i o n i s need-

ed t o make o p e r a t i o n a l d e c i s i o n s . These s i t u a t i o n s c o u l d be d i f f e r e n t

from those p r e v i o u s l y c o n s i d e r e d during planning. S i n c e the time s o l u -

tion method i s slow, the system operator has e i t h e r to overreact to

ensure system s e c u r i t y or t o make d e c i s i o n s that may put s t a b i l i t y of

the system at r i s k .

3. The power system o p e r a t i n g c o n d i t i o n s change d u r i n g the' course of the

day and year, while stability studies are done off-line for certain
 3

severe cases. This leads t o improper decisions i n some cases and,

t h e r e f o r e , may i n c r e a s e o p e r a t i o n a l e x p e n d i t u r e s .

From the above, i t i s c l e a r t h a t the power i n d u s t r y g r e a t l y needs an

alternative method to s o l v e the t r a n s i e n t stability problem. The a l t e r n a -

t i v e method cannot e n t i r e l y r e p l a c e the time s o l u t i o n , but i t should reduce

the number of s i m u l a t i o n s f o r each case and hence save a great d e a l of

computation time. A new method i s a l s o needed f o r system o p e r a t i o n s where

the system operator has o n l y minutes or hours to make an important opera-

tional decision. The best s o l u t i o n , of course, i s t o have an o n - l i n e method

that d e a l s w i t h system o p e r a t i o n s on a r e a l - t i m e b a s i s .

The d e s i r e d method f o r fast a n a l y s i s of t r a n s i e n t stability should

satisfy the f o l l o w i n g important requirements.

1. Provide fast and r e l i a b l e answers to the t r a n s i e n t stability problem

when a s p e c i f i e d disturbance i s given. The answer must particularly

i n d i c a t e whether the system i s s t a b l e o r n o t .

2. Include sufficient modelling o p t i o n s so t h a t the answer p r o v i d e d is in

the range of the a c t u a l system response.

3. Provide the n e c e s s a r y information about the degree of s t a b i l i t y and

system security so that the o p e r a t o r can make proper d e c i s i o n s to

ensure system security.

4. I f this fast method i s t o be used f o r r e a l - t i m e system o p e r a t i o n s , i t

should be adaptable to changes i n operating conditions, different

d i s t u r b a n c e s , and s t a b i l i t y controls.

Although e x t e n s i v e r e s e a r c h has been conducted i n this area, little

of the p r e v i o u s requirements have been a c h i e v e d so f a r . A great d e a l of

 4

research is still needed to e i t h e r improve the e x i s t i n g methods or propose

new methods i n order to f u l f i l l a l l these r e q u i r e m e n t s .

T h i s r e s e a r c h i s m o t i v a t e d by the c h a l l e n g i n g problem of d e v e l o p i n g a

satisfactory fast direct method f o r the assessment of t r a n s i e n t stability.

The main o b j e c t i v e s of t h i s t h e s i s a r e :

1. To develop a f a s t , reliable direct method f o r the assessment of t r a n s -

ient s t a b i l i t y of power systems.

2. To improve the power system model used by including a l l necessary

options to get as a c c u r a t e a system response as p o s s i b l e .

3. To predict the comprehensive transient stability region with security

boundaries f o r p o s s i b l e o n - l i n e assessment of t r a n s i e n t stability.

A review of the literature on the transient stability problem and

presently available solutions are g i v e n i n Chapter 2. I n Chapter 3 a new

d i r e c t method f o r s t a b i l i t y assessment i s introduced with preliminary appli-

cations on a simple one-machine infinite-bus power system. I n Chapter 4,

the technique is applied to multi-machine power systems using dynamic

equivalents and Taylor series expansions. Transient stability r e g i o n s are

also given by using the c a t a s t r o p h e theory. Chapter 5 p r e s e n t s the cap-

ability of the new method to include excitation, flux decay and system

damping i n the power system model. Chapter 6 summarizes the c o n c l u s i o n s and

achievements of t h i s project.
 5

CHAPTER 2

REVIEW OF STABILITY PROBLEM AND ITS SOLUTION

2*1 Introduction

The stability problem of power systems has been g i v e n new importance

since the famous blackout in Northeast U.S.A. in 1965. Considerable

research e f f o r t has gone i n t o the s t a b i l i t y i n v e s t i g a t i o n both f o r o f f - l i n e

and on-line purposes. It started a chain reaction affecting planning,

operation and control procedures of electric power systems. Since then

several studies have been conducted and new concepts and directions have

been suggested t o prevent i n s t a b i l i t y and ensure s e c u r i t y and r e l i a b i l i t y of

power systems.

In this chapter, a brief review of the stability problem is

p r e s e n t e d , f o l l o w e d by a l i t e r a t u r e review of the d i f f e r e n t methods used and

suggested to analyze i t .

2.2 The S t a b i l i t y Problem

A s t a b l e power system implies that a l l i t s interconnected generators

are o p e r a t i n g i n synchronism w i t h the network and w i t h each o t h e r . Problems

a r i s e when the g e n e r a t o r s o s c i l l a t e because of d i s t u r b a n c e s that occur from

t r a n s m i s s i o n f a u l t s on s w i t c h i n g o p e r a t i o n s .

There are two types of s t a b i l i t y problem i n power systems. Firstly,

the steady state stability problem which refers to the s t a b i l i t y of power

systems when a small d i s t u r b a n c e occurs in the power systems such as a

gradual change in loads, manual or automatic changes in excitation,

 6

irregularities in prime-mover input. Obviously these small disturbances

cannot cause l o s s of synchronism unless the system i s o p e r a t i n g a t , or very

near, i t s steady state s t a b i l i t y limit. This limit i s the g r e a t e s t power

that can be transmitted on a specified circuit, under certain operating

conditions i n the steady s t a t e , without loss of synchronism. The a n a l y s i s

of steady state s t a b i l i t y r e q u i r e s the s o l u t i o n of power flow equations and

swing equations over a period of a few miniutes. Governor and exciters

should a l s o be i n c l u d e d i n the steady s t a t e s t a b i l i t y analysis.

Secondly, t h e r e i s the t r a n s i e n t stability problem which a r i s e s from

a large disturbance i n the power system such as a sudden l o s s of generators

or l o a d s , s w i t c h i n g o p e r a t i o n s , or f a u l t s w i t h subsequent c i r c u i t isolation.

Such l a r g e d i s t u r b a n c e s c r e a t e a power unbalance between supply and demand

in the system. This unbalance takes place at the generator shafts and

causes the r o t o r s to o s c i l l a t e until a new steady state operating point i s

reached; or until the r o t o r s continue to oscillate and deviate from each

other and f i n a l l y some g e n e r a t o r s w i l l l o s e synchronism.

Loss of synchronism must be prevented or c o n t r o l l e d because i t has a

d i s t u r b i n g e f f e c t on v o l t a g e s , frequency and power, and i t may cause s e r i o u s

damage to g e n e r a t o r s , which are the most expensive elements i n power systems

[1]. The generators which tend to l o s e synchronism should be t r i p p e d , i . e .

disconnected from the system before any s e r i o u s damage occurs, and subse-

quently brought back to synchronism. While this can be done r e a d i l y with

gas and water turbine generators, steam turbine generators r e q u i r e many

hours to r e b u i l d steam and the operator has to shed l o a d to compensate f o r

the loss of g e n e r a t o r s . Loss of synchronism may a l s o cause some p r o t e c t i v e

 7

relays t o operate falsely and t r i p the c i r c u i t breakers of u n f a u l t e d l i n e s .

In such cases the problem becomes very complicated and may result i n more

generators l o s i n g synchronism.

2.3 B a s i c Power System S t a b i l i t y Model

The equation of motion of n g e n e r a t o r rotors i n a power system of

n-machines i s g i v e n by

M.6. + D. 6 = p - P = P , i = 1,2, . n (2.1)

i i ' i i i ei ai

where
= 2 ( G E E . cos 6 + B, . E . E . s i n 6 ) (2.2)
ei ij i J ij ij i j ij'
j=l

i n t e r n a l r o t o r angle of machine 2.

M. to 1^ = i n e r t i a constant,

moment of i n e r t i a ,

10 angular speed,

damping coefficient,

mechanical power i n p u t ,

e l e c t r i c a l power output,
ei
a c c e l e r a t i n g power.
ai
G ,B r e a l and imaginary p a r t s of reduced n o d a l admittance matrix.
ij* i j
6 6 - 6 .
ij i j
i n t e r n a l v o l t a g e s of machine i,j.
V E
j

Under steady state conditions, P .=0, and 6. i s c o n s t a n t . When a

ai i
{

disturbance occurs, P
. becomes d i f f e r e n t from zero and e q u a t i o n (2.1)
ai
describes the behaviour of 6 w i t h time. F o r machine i t o be s t a b l e , 6
 8

must settle to a constant v a l u e a g a i n and P . must r e t u r n to z e r o . For a

power system of n machines, n swing equations have to be s o l v e d i n order to

d e c i d e whether the power system i s s t a b l e or not when a s p e c i f i c disturbance

o c c u r s . . The s o l u t i o n of the swing e q u a t i o n s o b t a i n e d depends upon the model

of the power system elements such as generators (machine, exciter and

governor), transmission lines and loads. Furthermore, the choice of the

power system model w i l l depend upon the stability study to be carried out

and the p e r i o d of a n a l y s i s [ 2 ] .

2.3.1 Synchronous Generator

The most simplified model of the sychronous generator is the so-

called classical model. In this model the generator is described by a

constant voltage behind transient impedance. T h i s model i s a c c e p t a b l e for

the first swing transient stability a n a l y s i s which has a period of one

second or less [2]. The internal voltage i s constant i n magnitude only.

This representation neglects the effect of s a l i e n c y and assumes constant

flux linkages and a s m a l l change i n speed. The voltage equation i s given

by

|E| / 6 - V fc + r a I t + j x- l t
(2.3)

where |E|: magnitude of v o l t a g e behind transient impedance.

6: angle of |E| ( v a r i a b l e ) .

V terminal voltage.
t
r armature r e s i s t a n c e .
a
transient reactance.

I terminal current.
t
 9

When s a l i e n c y and changes i n f i e l d f l u x l i n k a g e s are taken i n t o account, i t

i s no l o n g e r p o s s i b l e to r e p r e s e n t the generator completely by an e q u i v a l e n t

circuit. The r e s u l t i n g equations have time-dependent c o e f f i c i e n t s unless

they a r e transformed by the use of Park's transformation into the s o - c a l l e d

direct- and q u a d r a t u r e - a x i s variables [2]. The model then involves both

differential and p h a s o r - a l g e b r a i c equations [ 3 ] . The d i f f e r e n t i a l equation

is:
i
dE ,
-1 =J - ( K - . - O (2.4)
dt ' f d 1

A
do
i

where E ^ i s t h e v o l t a g e p r o p o r t i o n a l to f i e l d flux l i n k a g e , T^ o i s the

direct-axis transient open-circuit time c o n s t a n t , E ^ i s the q u a d r a t u r e - a x i s

field voltage and E^. i s t h e v o l t a g e p r o p o r t i o n a l to f i e l d current. The

phasor equations of the above q u a n t i t i e s a r e :

E* = E - j ( x - x ' ) l , (2.5)
q q q d d

I t a t d d <}q

E = V + r I + jx I (2.7)
q t a t q t

where E ^ i s the v o l t a g e behind q u a d r a t u r e - a x i s reactance x^, 1^, 1^ a r e the

components of I f c a l o n g the d i r e c t - and quadrature-axes, respectively. Here

the machine angle 6 i s the angle of E ^ .

The above model i s d e s i r a b l e i n the t r a n s i e n t stability analysis f o r

the cases when a longer period of a n a l y s i s i s needed when a high-speed

voltage-regulator i s considered [ 2 ] .
 10

2.3.2 E x c i t e r and Governor C o n t r o l Systems

The exciter is a device that supplies flux to the synchronous

generator and directly controls the output voltage of the generator. The

exciter control system provides the proper field voltage to maintain a

desired system v o l t a g e during transient and steady s t a t e operations. There

are many types of exciter control systems i n use i n power systems [2]. The

basic components of an exciter control system are the regulator, amplifier

and exciter. The regulator measures the actual regulated voltage and

determines the v o l t a g e deviation. The deviation s i g n a l i s then a m p l i f i e d to

provide the signal required to control the exciter field current that

changes the exciter output voltage and hence results i n a new excitation

level for the generator. Reference [3] gives the detailed differential

e q u a t i o n s of the e x c i t e r c o n t r o l system.

The exciter control system must be Included in the transient

stability studies if the period of analysis is longer than one second.

Modern f a s t h i g h - c e i l i n g e x c i t e r s operate even w i t h i n a period of one second

and i f such an e x c i t e r i s c o n s i d e r e d then the e x c i t e r c o n t r o l system must be

included i n the g e n e r a t o r model.

The rotor speed of the synchronous generator is controlled by the

governor c o n t r o l system by v a r y i n g the m e c h a n i c a l power i n p u t . The governor

c o n t r o l system can be taken i n t o account i n the stability studies by solving

its differential equations simultaneously with the rest of the system

equations. I t i s necessary to i n c l u d e the governor e f f e c t i f the period of

analysis Is l o n g e r than one second [3].

 11

2.3.3 T r a n s m i s s i o n System and Loads

Transmission lines are u s u a l l y represented by nodal type equations.

C u r r e n t equations are w r i t t e n f o r each node i n the f o l l o w i n g manner:

n
1. = E Y E (2.8)
j-1 J J

where E . a r e nodal v o l t a g e s , Y. . a r e the c o e f f i c i e n t s of the i ' * 1

row of the
J i j

network admittance m a t r i x and 1^ i s the i n p u t c u r r e n t i n t o node i due to a

g e n e r a t o r or l o a d . The power i n p u t i n t o the network a t node i i s g i v e n by

P
ei-V iO E (2
- 9)

Loads are u s u a l l y r e p r e s e n t e d as c o n s t a n t complex power, constant impedance

or constant current at c o n s t a n t power factor. The equation f o r constant

complex power i s

S = P i + jQ = E li = constant (2.10)

f o r c o n s t a n t impedance i s
E
i
Z. = y~ = c o n s t a n t (2-11)
i
and f o r constant c u r r e n t a t c o n s t a n t power f a c t o r i s

Transmission l i n e s and l o a d s a f f e c t directly the s t a b i l i t y limits [4]. The

types of l o a d models mentioned above are not adequate models i f the load i s

of a dynamic type. I t i s very important to use more advanced l o a d models

 12

for dynamic loads especially when high voltage variations are expected

[5,6].

2.4 S o l u t i o n of the T r a n s i e n t Stability Problem

Transient stability r e f e r s to the amount of power that can be stably

transmitted when the power system i s subjected to a large disturbance such

as, faults, l o s s of l i n e s , l o s s of g e n e r a t o r s or loads and sudden changes i n

the tie-line flow. The main o b j e c t i v e of a t r a n s i e n t s t a b i l i t y study i s to

assess the possible loss of synchronism due to such large disturbances.

Transient stability studies enable power engineers to set up protective

devices and to design proper stability controls that ensure no loss of

synchronism.

The growth of large interconnected power systems demands careful

transient stability studies because some large disturbances could have

catastrophic results. On the other hand, rapid growth makes i t extremely

difficult to carry out such careful and detailed stability studies. An

easily reached conclusion is that an exact answer is difficult to obtain

[7].

There are three main approaches to solving the transient stability

problem. These approaches w i l l be discussed i n the f o l l o w i n g sections.

2.4.1. N u m e r i c a l I n t e g r a t i o n Methods

These methods solve the power system differential equations (swing

equations) during and after the transient period. From the response of
 13

the swing curves of the g e n e r a t o r s i n the power system an e x p e r i e n c e d power

e n g i n e e r can assess the s t a b i l i t y of the power system.

The numerical integration methods can be divided into two basic

types:

1. One-step methods: these methods r e q u i r e only information concerning the

previous time point to calculate the values at the next time point.

C o n s i d e r a s e t of n o n - l i n e a r differential equations

^ = f(y,t) (2.13)

where y r e p r e s e n t s a s e t of n v a r i a b l e s and f a v e c t o r of n f u n c t i o n s . The

one-step method c a l c u l a t e s a t p o i n t 1 s t a r t i n g from p o i n t 0 u s i n g a s t r a i g h t

l i n e defined by the d e r i v a t i v e a t p o i n t 0

yi = yo d r l o
+ h ( 2 a 4 )

where h i s the time i n t e r v a l , t ^ - t ^ . The new p o i n t can be used to c a l c u l a t e

the next one, and the i t e r a t i o n continues until the s o l u t i o n i s obtained.

These methods have h i g h accuracy when the time interval Is s m a l l , but the

accuracy decreases r a p i d l y as the time interval increases. Euler's method

and the step by step method g i v e n i n [1] a r e t y p i c a l one-step methods.

2. Multi-step methods: these methods require for each new point,

information about two or more preceding points. The predictor-corrector

method [8] c a l c u l a t e a p o i n t y as f o l l o w s
 14

*n+l * }
k=0
\Vk n + k=-l
b
k f
^
t
n-k ) ( 2
' 1 5 )

where appropriate values for the a and b constants are chosen for the

predictor and corrector formulas. f(y,t) indicates the derivative with

respect to time. Equation (2.15) i s the same as equation (2.14) f o r the

one-step method w i t h a =b =1, and a l l o t h e r c o n s t a n t s equal to z e r o .

r
n n ' n

The number s e r i e s approach [3] Is another type of m u l t i - s t e p method

where the generator angles versus time are obtained using the following

iterative equation

6 + - 6 + i ^ L + i ^ P _ (2.16)
n+2 n w a n+1

Equation (2.16) is derived from the swing equation by evaluating 6(t)

i n t e g r a t i o n by the t r a p e z o i d a l rule.

Other methods have been a p p l i e d to t h i s problem such as Runge-Kutta

method [10], an improved Gauss-Seidel approach [11], and the phase-plane

techniques [12]. The main advantage of m u l t i s t e p methods i s t h a t l o n g time

i n t e r v a l s can be used w i t h h i g h a c c u r a c y and stability.

In g e n e r a l , the n u m e r i c a l i n t e g r a t i o n methods are w i d e l y accepted and

used because of the f a c t t h a t these methods p r o v i d e d e t a i l e d r e s u l t s and are

capable of i n c l u d i n g d e t a i l e d models of power system elements.

The main disadvantage of the numerical integration methods for

transient stability studies are the formidable computation time required,

the need for human interpretation of the swing curves f o r assessing the

transient stability and the f a c t t h a t the method cannot be used for on-line

assessments of stability.
 15

In spite of the high speed computation of modern computers and

several refinements and developments of the technique [13-15] the digital

analysis of even a moderate s i z e d power system is still too time consuming.

Analog and hybrid simulation [16] have been suggested. In these schemes,

all the differential and algebric equations are s o l v e d i n analog form and

the rest of the work (initialization, s w i t c h i n g and logic operations) i s

taken care of by the d i g i t a l computer. But t h i s approach is s t i l l too slow

for on-line applications.

2.4.2 Lyapunov's D i r e c t Method

To improve on the slow computational speed of n u m e r i c a l integration

methods, Lyapunov's stability theory has been applied to the transient

stability problem.

Let the system be r e p r e s e n t e d by a s e t of d i f f e r e n t i a l equations

dX.
^JL = f (X 1 X ,
2 ... X , n t), i = l,...n. (2.17)

The origin of the state vector X is considered t o be stable, i . e .

f (0,0
i ... 0,t) = 0 , i = l,...n.

When the s e t of initial c o n d i t i o n s at the start of the d i s t u r b a n c e

are a l l known, Lyapunov's d i r e c t method w i l l determine, without finding the

actual trajectory of the s t a t e variables, i f the systm w i l l asymptotically

reach a stable point. T h e r e f o r e , much integration time can be saved as

stability can be determined directly from i n i t i a l conditions.;.

The method assumes a scalar, continuous Lyapunov function V(X,t)

exists i n the neighborhood of the o r i g i n such that,

 16

1. V(0,t) - 0

2. V(X,t) > 0 X e R, X * 0

3. fffidl <0 X e R
dt

where R i s a r e g i o n around the stable point X=0 and is called the region of

stability. If the above c o n d i t i o n s are valid for V(X,t), i t can be shown

that for a disturbance X e R (X * 0) then X e R as t -* fl0

. The main

difficulty here i s that the d e t e r m i n a t i o n of the boundary of the stability

region i s v e r y time consuming.

The a p p l i c a t i o n of t h i s method to the transient s t a b i l i t y problem of

power systems can be summarized i n the f o l l o w i n g steps [17-19].

1. I n i t i a l conditions are determined by a load flow study.

2. The admittance matrix of the power system i s reduced to the number of

generating buses for during-disturbance and post-disturbance

conditions.

3. The steady state equilibrium point of the post-disturbance system is

found by s o l v i n g the equation

P., - P = 0 , i = 1, ... n
i ei

where P^ and P ^
g are the mechanical input power and the electrical

output power, r e s p e c t i v e l y . The same e q u a t i o n has also to be s o l v e d to

find the closest unstable equilibrium point. Such a point helps to

define the r e g i o n of s t a b i l i t y around the stable state point.

4. The Lypanov's function V(X,t) is found and calculated. Thus the

stability regions f o r V=constant can be found.

 17

5. Integration of the differential equations for the during-disturbance

period. At each new point, the function V is c a l c u l a t e d with post-

disturbance conditions to determine whether the system i s s t a b l e when

the disturbance is cleared. The first instance of instability defines

the critical c l e a r i n g time f o r the disturbance.

Numerous methods have been suggested to d e r i v e a Lyapunov f u n c t i o n w i t h the

desired properties [20,21]. It should be noted here that different V

functions may yield different answers in the sense that they comprise

different subregions of the stability r e g i o n around the s t a b l e p o i n t [15].

The transient energy function method [22] is the most promising

approach among the direct methods. Remarkable progess has been made in

recent years [23,24]. However, the method i s s t i l l not suitable for on-line

a p p l i c a t i o n because of some p r a c t i c a l and computational d i f f i c u l t i e s [25].

The advantages which Lyapunov's direct method has over numerical

i n t e g r a t i o n methods are the following:

1. T h i s method i s c o m p u t a t i o n a l l y faster. Here the critical c l e a r i n g time

can be obtained i n a single i n t e g r a t i o n while the simulation technique

requires r e p e t i t i v e I n t e g r a t i o n f o r d i f f e r e n t assumed c l e a r i n g times.

2. For a given c l e a r i n g time, t h i s method w i l l i n d i c a t e whether the system

is stable or n o t . This s o l u t i o n i s d i r e c t and i t does not require the

integration of the system differential equations beyond the clearing

time.

3. It is shown that the value of the Lyapunov function and the critical

clearing time are related [26]. Hence, t h i s method can yield stability

margins or s t a b i l i t y i n d i c e s w i t h o u t making a c t u a l c a l c u l a t i o n s . Looked

 18

at from another a n g l e , t h i s method can i n d i c a t e whether a c l e a r i n g time

i s adequate or n o t .

Despite the above a t t r a c t i v e features, Lyapunov's d i r e c t method has

not received general acceptance from the u t i l i t i e s . The main reasons f o r

t h i s a r e the f o l l o w i n g p r a c t i c a l d i f f i c u l t i e s a s s o c i a t e d w i t h t h i s method:

1. Approximate mathematical model of the system; many simplifying

assumptions have been made i n a r r i v i n g a t a mathematical model of the

power system to make i t s u i t a b l e for i t s analysis by Lyapunov's d i r e c t

method. The critics of this method question the v a l i d i t y of the

simplifications which are necessary f o r construction of a proper

Lyapunov f u n c t i o n . Moreoever, the Lyapunov f u n c t i o n s a r e not unique and

several such functions can be c o n s t r u c t e d by u s i n g different dynamic

models of the power system.

2. The method i s conservative and, a l t h o u g h some s t a b l e point are r e a d i l y

identified, others are i n c o n c l u s i v e . Also, the method cannot predict

instability, therefore, i t may produce too many f a l s e alarms.

3. Determination of the r e g i o n of s t a b i l i t y ; from a p r a c t i c a l a p p l i c a t i o n

point of view, t h i s has been the most consuming p a r t of t h i s method. In

power systems, this requires determination of the u n s t a b l e equilibrium

points c l o s e s t to the post f a u l t s t a b l e e q u i l i b r i u m p o i n t s , and i n v o l v e s

the s o l u t i o n of n - n o n l i n e a r a l g e b r a i c equations.

2.4.3 Pattern Recognition Method *

The object of t h i s method i s t o determine a f u n c t i o n S(X) [27] such

that:
 19

> 0 for stable X

S(X) - {

< 0 for unstable X

S(X) is called the decision function and X i s the state variable , where

(X , X X^).

The classification procedure in the pattern recognition method is

shown i n F i g . (2.1) and summarized i n the following steps:

Tra i n i n g Feature Training

Data_* Decision
set extraction procedure

Figure (2.1)

1. Training Set:

In order to o b t a i n S(X), a t r a i n i n g set of various operating condi-

t i o n s must be a v a i l a b l e . Each o p e r a t i n g c o n d i t i o n or s t a t e i s s p e c i f i e d

by variables such as i n j e c t i o n powers, loads, generation powers, load

flows, voltage magnitudes and rotor angles. Each v a r i a b l e d e s c r i b i n g a

particular X condition constitutes a component of the pattern vector

X = (X^ Xj x
n ) * I d e a l l y , every conceivable pattern should be

included i n the t r a i n i n g s e t so t h a t i t c o v e r s the whole spectrum of the

system o p e r a t i n g conditions. Each p a t t e r n of the t r a i n i n g s e t must be

c l a s s i f i e d whether i t i s s t a b l e or unstable.
 20

2. Feature extraction:

The number of v a r i a b l e s obtainable from the t r a i n i n g s e t can be very

large. In pattern recognition, i t i s not d e s i r a b l e nor i s i t n e c e s s a r y

to use a l l the a v a i l a b l e variables to o b t a i n the c l a s s i f i e r function.

T h e r e f o r e , we have to choose the most Important v a r i a b l e s that contain

the necessary information about s t a b i l i t y . This can be done by p r a c t i -

cal experience with the power system or by u s i n g s t a t i s t i c a l measures.

The following function F [28] p r o v i d e s an easy and straightforward

measure of the i n f o r m a t i o n i n each v a r i a b l e .

where

m = mean of the v a r i a b l e i n the s t a b l e class.

s
m = mean of the v a r i a b l e i n theu n s t a b l e c l a s s .
u
cr = v a r i a n c e of the v a r i a b l e i n the s t a b l e c l a s s .
s
a = v a r i a n c e of the v a r i a b l e i n the u n s t a b l e c l a s s .
u
Feature extraction begins w i t h the computations of the F v a l u e s f o r

all the components of the p a t t e r n vectors I n the t r a i n i n g set. The

variable with the l a r g e s t F i s s e l e c t e d as the f i r s t feature. Then the

redundant information i s removed by c a l c u l a t i n g the c o r r e l a t i o n factor

between the f i r s t feature and the r e s t of the v a r i a b l e s , excluding

variables with high correlation factor with the f i r s t feature. The

procedure of c a l c u l a t i n g F w i t h the r e s t of the f e a t u r e s Is repeated and

s e l e c t i o n continues [27].
 21

3. T r a i n i n g Procedure:

Having selected the s i g n i f i c a n t features, the next step i s to o b t a i n

the d e c i s i o n f u n c t i o n S ( X ) . T h i s f u n c t i o n can be f i r s t o r d e r such as

S(X) = W + W. X, + ... + W X

o 1 1 mm

The w e i g h t i n g c o e f f i c i e n t s (W ,
Q ... W ) are determined such that
ffi

S(X) > 0 i f X i s s t a b l e

and S(X) < 0 i f X i s u n s t a b l e

There are many t r a i n i n g procedures a v a i l a b l e such as the l e a s t squares

method [29,30], linear programming [30,31] and an optimal search

algorithm [32].

Another possible classifier function i s of second o r d e r or q u a d r a t i c

form [27] g i v e n below

m m m
S(X) = W + Z W, Z. + Z Z W,. Z, Z,
1-1 1 1
1=1 j-1 i j 1 i

After the weights are determined, their performance are checked by

u s i n g them to c l a s s i f y the p a t t e r n s i n the t r a i n i n g s e t .

T h i s method has been improved [33-34], and i t i s found t h a t non-para-

m e t r i c approaches are more r e l i a b l e than the p a r a m e t r i c approaches mentioned

above.

The non-parametric approaches r e l y on e x p e r i e n c e to s e l e c t the proper

v a r i a b l e s to be used i n s t a b i l i t y c l a s s i f i c a t i o n [35].

The main advantages of p a t t e r n r e c o g n i t i o n a r e :

1. It i s suitable f o r o n - l i n e assessment of t r a n s i e n t stability if a reli-

a b l e d e c i s i o n f u n c t i o n i s determined.
 22

2. The method i s independent of the model of the power system that is

used.

The main d i s a d v a n t a g e s of p a t t e r n recognition are:

1. O f f - l i n e computations to o b t a i n the d e c i s i o n f u n c t i o n a r e excessive.

2. The decision function S(X) may very well be sensitive to system

configuration so t h a t a d i f f e r e n t S(X) i s needed f o r any change i n the

system configuration.

3. Correct classification of 100% cannot be achieved with large power

systems because i t i n v o l v e s u n l i m i t e d off-line computations.

2.5 Discussion of E x i s t i n g Methods

The growth of large interconnected power systems dictates the

necessity for on-line transient s t a b i l i t y assessment t o ensure the s e c u r i t y

of the power system and maintain r e l i a b l e service. The^ main problem of

l a r g e power systems i s t h a t i t i s i m p o s s i b l e to simulate exactly for on-line

purposes. Most direct methods provide global solution to the stability

problem with high a p p r o x i m a t i o n , except the method using the energy-type

Lyapunov function [24] which yields solution f o r the individual machine.

But w i t h t h i s method the problem Is how to d e f i n e the s t a b i l i t y region f o r

an i n d i v i d u a l machine and that only a s i m p l i f i e d power system model i s

used.

One may expect better results from the pattern recognition method

because of the f a c t that i t i s independent o f the model used t o s i m u l a t e the

power system. This method can be used for an on-line assessment of

stability even f o r i n d i v i d u a l machines, but i t cannot be g e n e r a l i z e d because

 23

each d e c i s i o n f u n c t i o n i s good o n l y f o r the system t r a i n e d f o r .

The motivation for this work i s to search f o r a method which i s

suitable f o r an o n - l i n e assessment of t r a n s i e n t stability that can be used

f o r every i n d i v i d u a l g e n e r a t o r i n the power system. I f such a method can be

found then the problem of transient stability can be d e a l t with on-line.

The most suitable method (from among the e x i s t i n g methods) f o r such an

application is pattern recognition because of the fact that i t is

independent of the systems e q u a t i o n s which couple the g e n e r a t o r s together.

2.4.1 A p p l i c a t i o n of P a t t e r n R e c o g n i t i o n to a Simple Power System

A recent t e c h n i q u e of p a t t e r n recognition [35] i s a p p l i e d t o a s i n g l e

machine-Infinite bus system shown I n F i g . 2.2. Two indices are used as

state variables, the k i n e t i c energy (K.E) and transmission margin (TM),

where

K.E = - MW2

and

TM = P - P
max c

M = i n e r t i a constant of the machine

W = speed

p = maximum power f o r p o s t - f a u l t c o n d i t i o n
max
P = l i n e power a t the i n s t a n t a f a u l t i s cleared
c

K.E and TM have o p p o s i t e c h a r a c t e r i s t i c s with loading conditions as

shown i n F i g . 2.3.

The d e c i s i o n f u n c t i o n S(X) i n terms of K.E. and TM can be w r i t t e n i n

the form
 j.52

j.16
0
j . 16

j.52

H - 3.0 T, Z T,
|E| 1.25 pu

2d j 0.1.6 pu

F i g . 2.2 One machine i n f i n i t e - b u s power system

TM(%)
KE(%) 1 0 0

90 110

80
70 105
60
50 100

70 80 90 100 (%)
Load

F i g . 2.3 C h a r a c t e r i s t i c s of KE and TM with load v a r i a t i o n

TM

KE

F i g . 2.4 The decision function f o r three d i f f e r e n t f a u l t locations

 25

S(.X) = W
x
z
+ <t> o
S(X) > 0 : stable

S(X) < 0 : unstable

where X = (K.E, TM), W i s c o e f f i c i e n t vector

<t> = constant
o

The technique i s applied t o the system shown i n F i g . 2.2 f o r three

different fault locations indicated. Three different decision functions

were determined f o r the t h r e e fault l o c a t i o n s as shown i n F i g . 2.4.

For f a u l t l o c a t i o n ( 1 ) , S i ( X ) = -TM + 20.71 K.E + 2.44.

For f a u l t location (2),S (X) 2 = -TM + 31.51 K.E 4 3.09

For fault location (3), S (X) 3 = -TM - 681.8 K.E + 56.1.

In conclusion, to r e p r e s e n t a l l possible fault l o c a t i o n s we have t o

build up a surface that consists of a l l p o s s i b l e fault locations. This

surface should contain a l a r g e number of l i n e s i n order t o cover the whole

r e g i o n of s t a b i l i t y . Therefore, the d e c i s i o n f u n c t i o n S ( X ) should represent

the e q u i l i b r i u m s u r f a c e of the s t a b i l i t y r e g i o n . T h i s c o n c l u s i o n brought up

the i d e a of u s i n g c a t a s t r o p h e theory t o d e f i n e the s t a b i l i t y r e g i o n .

 26

CHAPTER 3

APPLICATION OF CATASTROPHE THEORY TO POWER SYSTEM STABILITY

A preliminary a p p l i c a t i o n of the c a t a s t r o p h e theory, a mathematical

method, to the s t a b i l i t y assessment of power systems i s introduced i n this

chapter. The theory i s reviewed i n S e c t i o n 3.2 f o l l o w e d by a p p l i c a t i o n s t o

both steady state and transient stability of a simple one-machine

infinite-bus power system. This method provides a very good tool to

v i s u a l i z e steady s t a t e and t r a n s i e n t s t a b i l i t y regions.

3.1 Introduction

The sustained i n c r e a s i n g demand f o r e l e c t r i c a l power r e q u i r e s l a r g e r

interconnected power systems and operation at or near to full capacity.

Therefore, transient s t a b i l i t y of power systems becomes a major factor i n

planning and day-to-day operations and there i s a need for fast on-line

s o l u t i o n of t r a n s i e n t s t a b i l i t y to p r e d i c t any p o s s i b l e l o s s of synchronism

and to take the necessary measures to r e s t o r e stability.

Lyapunov's d i r e c t method and p a t t e r n r e c o g n i t i o n have been introduced

for fast assessment of t r a n s i e n t s t a b i l i t y and e v e n t u a l l y t o implement these

methods f o r o n - l i n e a p p l i c a t i o n s .

Catastrophe theory has been applied to the study of stability of

various dynamic systems [36] and in recent years to the steady state

stability problem of power systems by Sallam and Dineley [37]. An

attractive f e a t u r e of c a t a s t r o p h e theory i s that the s t a b i l i t y regions are

 27

defined i n terms of the system parameters bounded by the l i n e s of stability

limits.

The application of catastrophe theory i s extended to the transient

stability problem. A well defined transient stability region i n terms of

system parameters i s obtained; these regions are suitable for fast on-line

applications.

3.2 C a t a s t r o p h e Theory

It i s a n a t u r a l phenomenon t h a t sudden changes can occur as a r e s u l t

of smooth or gradual changes. Examples might include the breakdown of an

insulator when v o l t a g e is built up gradually, the collapse of a bridge by

gradual load increases and the l o s s of synchronism of g e n e r a t o r s i n a power

system when subject to smooth changes in operating conditions. The term

"catastrophe" is used for such sudden changes that are caused by smooth

alterations.

Catastrophe theory was originally presented by the French

mathematician Rene Thorn and published i n h i s book " S t r u c t u r a l S t a b i l i t y and

Morphogenesis" [38]. Thorn used differential topology to explain sudden

changes i n morphogenesis. The idea then attracted the attention of many

s c i e n t i s t s and has been a p p l i e d to a v a r i e t y of s c i e n c e s [36].

Catastrophe theory is a theory that explores the region of sudden

changes i n dynamic systems and deals with the properties of d i s c o n t i n u i t i e s

directly. I t has been d e f i n e d as the study from a q u a l i t a t i v e p o i n t of view

of the ways the s o l u t i o n s to d i f f e r e n t i a l e q u a t i o n s may change [39].

 28

Consider a system whose behaviour is usually smooth but which

sometimes (or i n some p l a c e s ) e x h i b i t s discontinuities. Suppose the system

has n state variables and c o n t r o l l e d by m independent v a r i a b l e s , and suppose

a smooth p o t e n t i a l function to describe the system dynamics exists. Given

such a system, what catastrophe theory tells us Is the following: The

number of qualitative different configurations of d i s c o n t i n u i t i e s that can

o c c u r depends not on the number of s t a t e v a r i a b l e s , which may be very large,

but on the number of control variables, which is generally small.

Particularly, i f the number of control variables i s not greater than four,

then there are only seven d i s t i n c t types of c a t a s t r o p h e s known as the seven

elementary catastrophes, and in none of these are more than two state

variables involved [40].

3.2.1. C a t a s t r o p h e Theory and Bifurcation Analysis

Consider a c o n t i n u o u s p o t e n t i a l f u n c t i o n V(X,C) which r e p r e s e n t s the

system behaviour, where X are the state variables and C are the control

variables. The potential function V(V,C) can be mapped in terms of its

control variables C to define the continuous region. Let the potential

function be represented as

V(X,C) : M X C * R (3.1)

where M, C are manifolds In the state space R and the c o n t r o l space R T

respectively.

Now we define the catastrophe manifold M as the equilibrium surface

that represents all critical points of V(X,C). I t i s the subset R Q

X R C

defined by

v x V (X)
C = o < - >
3 2
 29

where V ( X ) = V ( X , C ) and V
C x i s partial d e r i v a t i v e with respect t o X.

Equation (3.2) i s the s e t of a l l c r i t i c a l p o i n t s of the f u n c t i o n V(X,C).

Next we find the s i n g u l a r i t y s e t , S, which i s the subset of M t h a t

consists of a l l degenerate critical points of V. These a r e the p o i n t s a t

which

v v (x) = 0
x c

and V x V (X) =
C 0 (3.3)

The singularity s e t , S, i s then p r o j e c t e d down onto the c o n t r o l space

R r
by e l i m i n a t i n g the s t a t e v a r i a b l e s X u s i n g (3.3) and ( 3 . 2 ) , to o b t a i n the

bifurcation s e t , B. The b i f u r c a t i o n set provides a projection of the

stability r e g i o n of the f u n c t i o n V ( X , C ) , i . e . i t c o n t a i n s a l l non-degenerate

critical points of the f u n c t i o n V bounded by the degenerate c r i t i c a l point

at which the system e x h i b i t s sudden changes when i t i s subject to small

changes.

The seven elementary catastrophes of r<4 a r e l i s t e d i n Table (3.1).

The geometric a n a l y s i s of the c a t a s t r o p h e s that are used i n t h i s t h e s i s a r e

presented i n detail i n Appendix (Al). A simplified a n a l y s i s of the seven

elementary c a t a s t r o p h e s Is g i v e n i n Reference [ 4 0 ] .

3.3 A p p l i c a t i o n s t o the Steady S t a t e S t a b i l i t y Problem

The a p p l i c a t i o n of c a t a s t r o p h e theory t o the steady state stability

of power systems was first introduced i n Reference [37]. However, that

application was limited only t o s a l i e n t - p o l e type synchronous generators.

When the same procedure was applied to cylindrical-rotor generators,

Catastrophe C o n t r o l Space State Space Function Catastrophe Manifold
Dimension Dimension

2
Fold 1 1 1/3 x - a x
3

x -a

3 .
Cusp 2 1 x^-ax-1/2 b x 2
x -a-bx

4 , 2
Swallowtail 3 1 1/5 x - a x - l / 2 b x - l / 3 c x
5 2 3
x -a-bx-cx

5 2 3
Butterfly 4 1 1/6 x - a x - l / 2 b x - l / 3
6 2
cx -dx
3 4
x -a-bx-cx -dx

3x +a+cy
3 3
Hyberbolit 3 2 x +y +ax+by+cxy

3y +b+cx
2

2 2
3x -y +a+2cx

3 2 2 2
Eliptic 3 2 x -xy +ax+by+cx +cy -2xy+b+2cy

2xy+a+2cx

2 4 2 2
Parabolic 4 2 x y+y +ax+by+cx +dy x +4y +b+2dy
2 3

T a b l e 3.1 The Seven-Elementary C a t a s t r o p h e s

 31

unjustified answers were obtained [37] . In t h i s s e c t i o n , a new procedure

f o r a p p l y i n g c a t a s t r o p h e theory to the steady state s t a b i l i t y of c y l i n d r i c a l

rotor generators i s presented using Taylor series expansion around the

initial steady state operating point. This s e c t i o n i s a complement t o the

work done by Sallam and D i n e l e y and i t i n t r o d u c e s a g e n e r a l procedure which

is a p p l i c a b l e to a l l types of g e n e r a t o r s . Furthermore, the damping and

governor e f f e c t s can a l s o be i n c l u d e d u s i n g t h i s method.

3.3.1 C y l i n d r i c a l - r o t o r I n f i n i t e - b u s Power System

Consider the one-machine i n f i n i t e - b u s power system of F i g . 3 . 1 . The

steady s t a t e output power P g i s g i v e n by

P = P sin6 (3.4)
e m

where P =
m X,
d

S -is--the machine i n t e r n a l v o l t a g e , V i s the i n f i n i t e - b u s v o l t a g e , i s the

mchine reactance and 6 i s the r o t o r angle (angle of i n t e r n a l voltage E ) .

The power angle curve i s g i v e n i n F i g . 3 . 2 .

Suppose t h e i n p u t power to the machine i s i n c r e a s e d smoothly from P Q

to P^. The machine w i l l o s c i l l a t e between 6Q and 6 2

a s
shown In F i g . 3 . 2 .

If the new operating angle i s higher than the maximum power limit, the

machine w i l l lose synchronism.

The energy of o s c i l l a t i o n s between 6 and 6^ i s g i v e n by

6
2
f Pa d6 * (3.5)
6
o
 32

jo.72

g infint
P bus

E = 1.71 V = 1. /o.o
6 = 36.5 jo.72
o

F i g . 3.1

6
0 6
1 6
2
Angle

F i g . 3.2 The power-angle curve

 33

P = P - P sin 6 (3.6)
a i m

P i s the a c c e l e r a t i n g power.

We now expand equation (3.6) by T a y l o r s e r i e s around 6^ a t t=0, P a

becomes

P ( 2 )
t 2

P (t) = P
a a
( 0 )
+ P ( 1

a
>t + J-
j
+ ...
2
(3.7)

where

(m) _ a
(3.8)
, m
d t o

Therefore

P ( 0 )
- P - P sin 6
a i m i

d P
(1) _ P cos 6 6=0
m o
d t

(1) (3.9)
d P
(2) P sin 6 6 P cos 6 6
m o m o
d t

= -P cos 6 6
m 0
 34

(2)
/Q\ D P

(3) _ a
= P cos 6 6 J
+ p sin 6 (26)6
m o m o
d t

+ P sin 6 6 6 - p cos 6 6=0

m o m o

,(4) _ a
- - P cos 6 N
,
6 ' * + 3P sin 6. 6'
a
" dt m 0 m 0
0

where 6 = 0 at t = 0

Since the change i n the power input of the machine i s smooth and

small, the Taylor series expansion determines the exact region of

oscillation, from 6 ^ t o 6 ^ . The a c c e l e r a t i n g power i s then g i v e n by

p (2) 2 t p (2) 4 t p (6) 6 t

P (t) = P ( 0 )
+ a
+ -5 + (3.10)
a ' v
a 2! 4! 6!

Let 6 = 6 Q + \ Y t ,
2
( 6 - 6 Q ) ^ y t 2
, 6' - \ Y t 2

,(0)
where Y
M

Put 6 - 6 Q = 6 ' =^ y t and s u b s t i t u t e i n e q u a t i o n (3.10) t o o b t a i n

,(6)
P<>6. P ( 4 )

+ 'J- ,3
P (6') - p ( 0 )
: + b l
+ 6 (3.11)

6Y 18 Y

6' r e p r e s e n t s the change i n r o t o r angle corresponding to changes i n

power i . e . a t t = 0 , 6' =0

Evaluating equation (3.5) t o o b t a i n the energy f u n c t i o n form 6 ^ to

 35

6^, we o b t a i n the energy f u n c t i o n ,

p (l) ,2 6 p (2) ,3 6 p (3) ,4 6

V = P ( }
6' + .1 + _? + (3.12)

a
2 3! 4!

Equation (3.12) represents a catastrophe cusp equation, see Table

(3.1). The first derivative of the energy f u n c t i o n (3.12) t h a t represents

the catastrophe manifold ( e q u i l i b r i u m s u r f a c e ) of a l l o p e r a t i n g p o i n t s of

the power system i s g i v e n by

V v = P > (
+ P ( 1 )
6' + P < >iZ
2
+ !f (3.13)
6' a a a
2! 3!

To put e q u a t i o n (3.12) i n the cusp c a t a s t r o p h e manifold, the second order

term must be e l i m i n a t e d .

Let 6' = X - a

P <> 2

and a = a
(3.14)
p (3)
a

Substitute equation (3.14) i n (3.13) to o b t a i n

2
p (2) p O)
V V = (P ( 0 )
+ K ) + (P ( 1 )
- ) X + JL X 3
- 0
X a v
a' a ,
2P ( 3 ) 6

p (2) 2

= X 3
+ (P ( 1 )
- X + -J (P^ 0 )
+ K ) - 0
a

p (3) a
2 p (3) p (3)
a a a
 36

X 3
+ aX + b = 0 (3.15)

where

p (2) 2

a = _ i _ ( P CD - I f )
p O) 3
2 p (3)
a a

b - _1_ ( p <>
a + K )
p (3) 8

P () 2
P O) a 3

" * 2! 3!

Equation (3.15) i s e x a c t l y t h e cusp c a t a s t r o p h e m a n i f o l d ( s e e Appendix B ) .

The steady state stability region i s calculated from e q u a t i o n (3.15) by

projecting i t down onto the c o n t r o l space ( a , b ) , where:

V 2
V = 3 X 2
+ a = 0 (3.16)
A.

From e q u a t i o n (3.16) and e q u a t i o n (3.15) the s t a t e v a r i a b l e X c a n now

be e l i m i n a t e d t o get the b i f u r c a t i o n s e t B,

4 a 3
+ 27 b 2
= 0 (3.17)

Equation (3.17) r e p r e s e n t s the steady state stability limits of the power

system i n terms of the c o n t r o l parameters a and b as shown i n F i g . 3.3.

 37

Fig. 3.3 The b i f u r c a t i o n s e t of the cusp c a t a s t r o p h e which r e p r e s e n t s

the steady s t a t e s t a b i l i t y r e g i o n .
 38

3.3.2 Numerical Results

Consider the one-machine i n f i n i t e - b u s system of F i g . 3.1, g i v e n the

following data

H = 3 s , V ro = 1. /0.0 pu

E = 1.71 pu , 6 = 36.5
q pu

X, = 1.05 pu , X_ = 0.36 pu
d t

Following the procedure i n the p r e v i o u s s e c t i o n , the s o l u t i o n of equation

(3.15) w i l l g i v e the new o p e r a t i n g angle f o r any change i n the i n p u t power.

The steady state s t a b i l i t y region i s shown I n F i g . 3.3. F o r any l o a d i n g

condition, the l o c a t i o n of the c o n t r o l parameters (a,b) on the b i f u r c a t i o n

set of F i g . 3.3 w i l l determine the steady state s t a b i l i t y of the system.

The r e s u l t s o b t a i n e d i n F i g . 3.3 agree w i t h the power e q u a t i o n (3.4).

It should be clear that the change i n the l o a d i n g c o n d i t i o n must be s m a l l

and w i t h i n the d e t e r m i n i n g r e g i o n of the T a y l o r s e r i e s order used. However,

a higher order Taylor series can be used with a suitable higher

catastrophe.

This method can be extended f u r t h e r to Include the speed governor

effect.

3.4 A p p l i c a t i o n t o the T r a n s i e n t S t a b i l i t y Problem

The application of the catastrophe theory to the steady state

stability problem is fairly straightforward, as shown i n the previous

section. I n the case of t r a n s i e n t stability, the s i t u a t i o n i s different

because there are two switchings (discontinuities) during the transient

 39

>

>

. 3.4 The energy function s t a b i l i t y criteria

a. stable
b. c r i t i c a l l y stable
c unstable
 40

p e r i o d , one at f a u l t occurrence and the o t h e r a t f a u l t c l e a r a n c e . Before we

attempt to apply the catastrophe theory we need to find a continuous

function that represents the system behaviour during the transient period.

We need a l s o to d e f i n e the degenerate and non-degenerate c r i t i c a l points i n

terms of t r a n s i e n t stability.

C o n s i d e r a g a i n the one-machine i n f i n i t e - b u s system of F i g . 3.1. If a

fault occurs on one of the lines near the machine bus, the rotor will

a c c e l e r a t e and g a i n k i n e t i c energy. I f the f a u l t i s c l e a r e d a t the critical

c l e a r i n g time, the k i n e t i c energy generated by the f a u l t w i l l be absorbed by

the system and the gained energy at the end of the t r a n s i e n t p e r i o d w i l l be

exactly zero; the system i s c o n s i d e r e d to be critically stable. A typical

energy curve for a fault c l e a r e d at different clearing times i s shown i n

Fig. 3.4. In terms of catastrophe theory, curve (b) of F i g . 3.4 can be

considered as the energy e q u i l i b r i u m surface or catastrophe manifold at

which the kinetic energy equals the potential energy of the system. All

non-degenerate critical points l i e on the energy e q u i l i b r i u m s u r f a c e which

corresponds to critical clearing times. The degenerate critical points

which are d e f i n e d by the b i f u r c a t i o n s e t , are the p o i n t s which correspond to

the transient stability limits of the power system at which any small

d i s t u r b a n c e of the power system w i l l d r i v e the system u n s t a b l e r e g a r d l e s s of

the clearing time.

I n summary, s i n c e the energy f u n c t i o n d u r i n g the t r a n s i e n t period i s

continuous and represents the power system behaviour, the 'energy balance

equation will represent the e q u i l i b r i u m surface that decides the critical

 41

clearing time and the t r a n s i e n t stability limits define the degenerate

c r i t i c a l points.

34.i. Single-Machine I n f i n i t e - b u s Power System [42]

Consider the power system of F i g . 3.5 which c o n s i s t s of one machine,

an infinite bus and two t r a n s m i s s i o n lines ab and dc. Representing the

synchronous machine by a c o n s t a n t voltage source behind a reactance (the

classical model [ 2 ] ) , the swing e q u a t i o n r e p r e s e n t i n g the system behaviour

is g i v e n by

.2,
M = P. - P - P (3.18)
, i e a
dt z

where P = P s i n 6' I s the e l e c t r i c a l power output,

e max

M = inertia c o n s t a n t of the machine.

. > P^ = mechanical i n p u t power ( c o n s t a n t d u r i n g t r a n s i e n t ) .

P^ = a c c e l e r a t i n g power.

6 = rotor angle.

Consider a three-phase fault on line ab cleared a t the critical

clearing time t . The accelerating power P will exhibit two

c a

discontinuities, one when the f a u l t occurs and the other when the f a u l t i s

cleared. Multiply equation (3.18) by 6 and i n t e g r a t e w i t h r e s p e c t t o time

u s i n g the p o s t - f a u l t network c o n d i t i o n s to o b t a i n

1 M 6 = P cos 6 + P, 6
2
- P cos 6 - P 6 (3.19)
2 c m c l e m m i m ?

where
 42

6^ = c r i t i c a l clearing angle.

6^ = w = speed a t c r i t i c a l clearing.

= maximum power of p o s t - f a u l t network.

6^ = u n s t a b l e e q u i l i b r i u m angle (maximum a n g l e ) .

The L.H.S. of e q u a t i o n (3.19) r e p r e s e n t s the k i n e t i c energy (K.E)

generated d u r i n g the f a u l t - o n p e r i o d and the R.H.S. r e p r e s e n t s the p o t e n t i a l

energy (P.E) of the p o s t - f a u l t network. I f the f a u l t i s c l e a r e d a t the

c r i t i c a l c l e a r i n g time, then

K.E = P.E

or K.E - P.E = 0 (3.20)

and for stability

K.E < P.E (3.21)

Thus equation (3.19) r e p r e s e n t s the e q u i l i b r i u m s u r f a c e f o r the t r a n s i e n t

stability or i n terms of c a t a s t r o p h e theory i t represents the c a t a s t r o p h e

m a n i f o l d , N, which i s the g r a d i e n t o f the energy i n t e g r a l f u n c t i o n V ( 6 ^ ) .

N = V. V(6 ) =
o c ' _ c m
1
M 6 - P cos 6 - P, 6
c i c
2
+ P
m
cos 6 + P, 6
m i m
=0
c 2
(3.22)

We define also the s i n g u l a r i t y s e t , S, as the s e t of t r a n s i e n t stability

l i m i t s of the power system as

V 2
V(6 ) = 0 (3.23)
o c
c
U s i n g T a y l o r s e r i e s expansion t o approximate 6 and 6 as a f u n c t i o n of time
c c

(which has been r e p o r t e d to be a v e r y good approximation f o r the f i r s t swing

transient s t a b i l i t y a n a l y s i s [43]) we get

 43

6 = W = Yt and 6 =6 +1 yt 1
(3.24)
c c c c < > 2 c

where Y = a c c e l e r a t i o n a t the i n s t a n t of f a u l t occurrence

Y = I [P. - P ( t )] (3.25)
M

Let X = I Y t 2
(3.26)
2 C

and K = p 6 + p cos 6
mm m m

Substituting f o r 6^ a n d 6^ by e q u a t i o n s (3.24) and (3.26), r e p l a c i n g the

cosine term (cos 6^) by i t s expansion, then e q u a t i o n (3.22) becomes

(6 + X ) 2
(6 + X ) 4

N = V R V ( 6 ) = MYX - P [l- - + ] - P f 6 + X ) + K = 0
6
c c m a
. 21 4! 1

(3.27)

T r u n c a t i n g the c o s i n e expansion t o the f o u r t h order we get

(6 + X ) 2
(6 + X ) 4

N = MYX - P +2 P - P -P,(6+X)+K=0 (3.28)

' ma 2 ! ma ^, m a i o '

S i n c e e q u a t i o n (3.28) I s a f u n c t i o n of c l e a r i n g time which i s u s u a l l y small,

thus the fourth-order truncation would be enough to represent the

equilibrium s u r f a c e N.
 44

N = - J l i x 4
- J ^ 6 x 3
+ [2-) P X + (MY + 6 P - J
2
63
- P J X
24 6 4 o ma 6 o i '

+ (J5i 6 2
- 6 4
- P - P fi + K) = 0
2 0
24 0 1113

P P
Let A = , B = _S. 6
24 6

2-6 2

C = ( _) P , D = (MY + 6 P - 6 J
- P )
v y
m a oma o i
4 t
6

and E = JUi 6 - 2
6 4
- P - P 6 + K)
o 2 24 1113

N = -AX - BX + CX + DX + E = 0
4 3 2
(3.29)

Equation (3.29) i s structurally stable and i n the form of the

swallowtail catastrophe (Appendix A) except we need to eliminate the cubic

term.

Let X = y - a

N = -A(y-a) 4
- B(y-a) 3
+ C(y-a) 2
+ D(y-a) + E = 0

= -Ay + (4A<x - B ) y + (C-6Aa + 3Ba)y

4 3 2 2

+ (D - 2ca - 3Ba 2
+ 4Aa )y - A a + Bot + Cot - Dot + E = 0
3 4 3 2

IB
then a=
4 A
Substitute f o r ot and get
 45

8 A 2 B 8 .2
A

4 A 16 2 A 256 3 A

d i v i d e by -A

4 rC 3 B 2
i 2 rD 1 CB 1 B^ \
N = y - I- . Jy - I- Jy
A 8 .2 A 2 .2 8.3

2 2 4
_ E + D * _ - _ _ - _1_ 5_ = 0
1 4
A 2 1 6
A 3 2 5 6
A 4

Let

u = - + !-)
A 8 2 A

v= -(E-Ic!--!^) (3.30)
A 2 . 2 8.3
A A

2 4
E . D B C B 3 B
and w = - +
4
A 2 1 6
A 3 2 5 6
A 4

then N becomes

N = y 4
+ uy 4
+ vy + w = 0 (3.31)

Equation (3.31) i s the m a n i f o l d of the s w a l l o w t a i l c a t a s t r o p h e

 46

1 5 1 3 1 2
V(y) = y + - uy J
+ vy^ + wy (3.32)

5 3 2

We need to find the singularity s e t , S, which i s the subset of N that

consists of a l l the degenerate c r i t i c a l p o i n t s of V. These are the p o i n t s

at which
V V(y) = M = 0
y

and V y
2
V(y) = 0 (3.33)

V y
2
V(y) = 4y 3
+ 2uy + u = 0 (3.34)

It i s interesting t o note that the control variable (U) i n this case i s

c o n s t a n t and n e g a t i v e .

3B 2

We have u = - ( + )
A 8.2
A

P P
i. 1 1 1 3
T> ma c
where A = , B = o
24 6

2 - 6 2

and C = P (-
ma

thus,

u = - (12 - 66 2
+ 66 ) = - 12
2

o o

T h i s r e s u l t reduces the b i f u r c a t i o n m a n i f o l d from three dimensions to

only two dimensions i n u and W. We can plot the b i f u r c a t i o n s e t of the

F i g . 3.6 The t r a n s i e n t s t a b i l i t y l i m i t s g i v e n by the s w a l l o w t a i l
c a t a s t r o p h e f o r the power system of F i g . 3.5
 48

power system u s i n g e q u a t i o n s (3.31) and (3.34) (see Appendix A f o r d e t a i l s ) .

The boundaries of the b i f u r c a t i o n s e t of F i g . 3.6 represent the degenerate

transient stability limits of the power system.

It should be noted that f o r a g e n e r a t o r the c o n t r o l v a r i a b l e s u and W

are p o s i t i v e and f o r a motor they are n e g a t i v e .

The region of transient stability of the power system has to be

i n s i d e the p o s i t i v e s i d e of F i g . 3.6 and the f o l l o w i n g c o n d i t i o n s have to be

met:

1 2
Since X = _ Yt > 0

2 c

where t i s the c r i t i c a l clearing time

and X = y - a

then y - a > 0 (3.35)

or y > a (3.36)

IB c
we have a = = o

4 A

hence, y > 6^ (3.37)

For the s t a b i l i t y region bounded by e q u a t i o n (3.31), (3.34) and the

constraints y > 6 Q e q u a t i o n (3.31) has f o u r c r i t i c a l p o i n t s , two maxima and

two minima. The c r i t i c a l c l e a r i n g time i s r e p r e s e n t e d by the f i r s t positive

critical p o i n t of y > 6.
q

3.4.2 Results

The transient stability region of the system shown i n F i g . 3.5 i s

defined by equations (3.31), (3.34) and (3.37). Changing the fault

49
 50

Fig. 3.8 The t r a n s i e n t s t a b i l i t y r e g i o n i n terms of stability limits

and c r i t i c a l c l e a r i n g times.
 51

location, and f o r e a c h f a u l t the l o a d i n g c o n d i t i o n (P^) i s changed from .1

to 1.2 pu. The shaded areas of F i g u r e 3.7 represent the r e g i o n of t r a n s i e n t

stability for a l l possible fault l o c a t i o n s and loading conditions. A more

comprehensive view of the s t a b i l i t y r e g i o n can be v i s u a l i z e d by p l o t t i n g the

region in three dimensions where y is the singularity (y > a > 0)

representing the c r i t i c a l c l e a r i n g time

T c - /
2 ( y
" C )
(3.38)
Y

Figure 3.8 shows the transient s t a b i l i t y region i n three dimensions

v. W and t . The third dimension t i s c a l c u l a t e d from the catastrophe

' c c r

manifold and p l o t t e d over the shaded arc of F i g u r e 3.7. This q u a l i t a t i v e

view of the stability region cannot be achieved by other d i r e c t methods of

transient stability, because these other methods use state variables to

define the region of s t a b i l i t y which depends on operating conditions, fault

locations and time. But by using the catastrophe theory, the stability

region i s defined by two c o n t r o l v a r i a b l e s , u and W, only and includes a l l

possible fault l o c a t i o n s and loading conditions. The control variables u

and W can be written as a function of operating conditions and fault

l o c a t i o n s by u s i n g e q u a t i o n s (3.30) and s u b s t i t u t i n g f o r A, B, C, D and E to

get

u = 2L (My + Pj + 8(6 3
- 6 ) (3-39)
p
ma r

w = 24(6
o
2
+ til
p

ma
o p

ma
6
o
+ *L
p

ma
6 - JL + 1)

In fact the region of s t a b i l i t y d e f i n e d by u, W and T i s fixed and

 52

the machine i s s t a b l e when i t i s o p e r a t i n g Inside the r e g i o n . These r e s u l t s

are I d e n t i c a l to that of the time s o l u t i o n .

3.4.3 Equivalent Two Machine Power System [44]

Consider the power system of F i g . 3.9 which c o n s i s t s of three power

plants A, B and C. A three-phase f a u l t i s a p p l i e d on line 5-6 near bus 6.

It i s found t h a t machines B and C are swinging c o h e r e n t l y against machine A.

In this case B and C can be combined together to form a l a r g e machine B+C,

and the system can be reduced to a two-machine power system. The swing

e q u a t i o n of machine A a g a i n s t the e q u i v a l e n t machine (B+C) can be w r i t t e n i n

the form

,2c
M - = P - (P + P s i n (6 - a)) (3.40)
7 I c m

dt Z

The same procedure of the previous section can be applied to

calculate the equilibrium surface from the energy balance equation [44].

Similar transient s t a b i l i t y regions of F i g s . 3.7 and 3.8 are also obtained

in this case.

The critical c l e a r i n g time c a l c u l a t e d from the catastrophe manifold

equation for t h i s case i s .348 seconds and that c a l c u l a t e d by time s o l u t i o n

is .35 seconds. The method shows very close agreement w i t h the numerical

integration method. When e i t h e r the fault l o c a t i o n or the loading condi-

tions are changed, the machines may respond to the disturbance in different

coherency. This dictates d i f f e r e n t equivalence calculations for different

fault locations and loading conditions. Therefore, a faster method

F i g . 3.9 The equivalent two machine power system
 54

is needed to assess the response of the system to d i f f e r e n t contingencies.

This problem w i l l be addressed i n the next chapter.

3.5 Summary

The preliminary applications of catastrophe theory to the transient

stability problems of simple power systems have shown three attractive

advantages:

1. The transient stability region is well defined i n terms of the control

variables (system parameters) r e g a r d l e s s of the state variables. The

stability r e g i o n s are bounded by the stability l i m i t s which p r o v i d e very

good insight to the security boundaries. Adequate stability controls

can be also designed and applied according to the transient stability

limits.

2. Extremely few computations are needed to define the stability region,

each transient case can be easily calculated in terms of the system

parameters without r e p e t i t i v e i t e r a t i o n s .

3. The method i s a serious candidate to be used f o r o n - l i n e assessment of

transient stability, security analysis and for the application of

transient stability controls such as dynamic breaking, fast valving,

etc.
 55

CHAPTER 4

TRANSIENT STABILITY REGIONS OF MULTIMACHINE POWER SYSTEMS

4.1 Introduction

The transient stability problem of multi-machine power systems is

much more c o m p l i c a t e d than the simple power systems analyzed i n Chapter 3.

The analysis in this case involves every machine in the power system

without equivalencing any machine. The swing equation of each machine

depends upon the movements of the o t h e r machines coupled t o i t .

There have been some major difficulties associated with the

application of direct methods to the transient stability problem of

multi-machine power systems. Although n o t i c e a b l e p r o g r e s s has been r e p o r t e d

in dealing with these difficulties [45-48], the challenge of fast on-line

transient stability assessment is s t i l l not overcome. These c h a l l e n g i n g

problems are [49]:

i. Power system model: a p p l i c a t i o n s of d i r e c t methods are l i m i t e d to the

classical model of power systems. In this model, generators are

represented by constant v o l t a g e magnitude behind transient reactance,

loads are represented by constant impedances, transfer conductances

are usually neglected or approximated, although they are significant

network elements [ 5 0 ] .

ii. Although fast exciters respond within the first swing period (.1

second), none of the existing direct methods consider excitation

effects.
 56

iii. In practice, power engineers are usually not interested in the

critical c l e a r i n g time (which i s the main g o a l of d i r e c t methods) but

rather i n the amount of power that can be d e l i v e r e d without risking

system s e c u r i t y f o r s p e c i f i e d c l e a r i n g times.

iv. Network r e d u c t i o n and c a l c u l a t i o n s of s t a b l e and unstable equilibrium

points for large power systems are time consuming. For unstable

equilibrium points, existing algorithms cannot always be relied upon

to converge to the r i g h t s o l u t i o n .

Two methods w i l l be introduced i n this chapter to p r o v i d e an energy

function suitable for the a p p l i c a t i o n of catastrophe theory, namely the

c r i t i c a l machines dynamic e q u i v a l e n t method and a method u s i n g T a y l o r s e r i e s

expansion of the accelerating power during the transient period. Both

methods need the identification of the critical machines for each

disturbance considerd.

The transient stability regions are then defined in terms of the

system parameters which enable power engineers to define the security

r e g i o n s and apply p r e v e n t a t i v e c o n t r o l methods.

The difficulties mentioned above w i l l be dealt with i n the proposed

methods. Transfer conductance will be fully Included; damping and

e x c i t a t i o n response w i l l be a l s o i n c l u d e d and discussed i n Chapter 5.

4.2 Dynamic E q u i v a l e n t of the C r i t i c a l Machines

The equation of motion of machine i i n a multi-machinfe power system,

using c l a s s i c a l model r e p r e s e n t a t i o n i s g i v e n by
 57

- P (4.1)
mi ei

n
where P (4.2)
ei

= e l e c t r i c a l power output of machines

P
mi =
mechanical power i n p u t

= intertia constant

6 = r o t o r angle

= speed d e v i a t i o n

= internal voltage

g^j = transfer conductance

b . = transfer susceptance

When a fault occurs in a large power system only a few machines

ttct.iv?.i-.y reBpon.se to the f a u l t and tend to l o s e synchronism. These machines

are known as the critical machines [45,47]. Therefore, i t i s enough to

study the behaviour of the c r i t i c a l machines w i t h r e s p e c t to the r e s t of the

power system i n order to e v a l u a t e the transient s t a b i l i t y of the system f o r

a specific fault.

Consider that machine k i s the critical machine(s) for a specific

disturbance. T h i s machine i s c o n s i d e r e d to be o s c i l l a t i n g a g a i n s t the rest

of the power system which i s not significantly affected by the disturbance

and w i l l be c o n s i d e r e d moving as one machine

n
Let M, (4-3)
0
 58

1 n

U
M Q i*k 1 1

Mg and 6 Q a r e , r e s p e c t i v e l y , the i n e r t i a constant and t h e a n g l e of the

c e n t e r of angle of the power system e x c l u d i n g the c r i t i c a l machine.

Let e k = 6 k - 6 0

- 6, - i _ I ( M l )
K
M Q i*k 1 1

also

k
* M Q i*k 1

9 . - 6 . - 1 - S (M 6 ) (4.5)
k
* M Q i*k 1 1

We s u b s t i t u t e equation (4.1) i n t o (4.5) t o o b t a i n the swing e q u a t i o n of

machine k w i t h r e s p e c t t o the c e n t e r of angle (COA).

\~V^-r- J P
mk " e k

\
P
1

M
o
*
i
n

,
(

t
P

k
mi- ei) P
<'>
4 6

V k " mk '
P P
ek " ^ X < mi " e i >
P P
<'> 4 7

M i*k
Q

Substituting f o r P ^ a n d P ^ from e q u a t i o n g (4.2) and s e p a r a t i n g 9^

from the r e s t of the system we get the swing equation of the c r i t i c a l

machine k a g a i n s t the r e s t of the power system. This i s explained i n the

following steps: >

 59

F i r s t we l e t <= E g ^

and C^-^y^
E q u a t i o n (4.7) becomes

"kV P
mk " kk - X
D
< ij D C O S
kj +
Scj S i n
V

M, n n
- JL Z (P - Z (D. cos 0 + C, , s i n 9 ) j
J-l
m i
M Q i*k l j i j l j l j

Separate the term B ^ k from the l a s t term and o b t a i n

Vk - mk P
" kk " ^
D
X ^ mi " jP
k < ijD C O S 0
i j + C
i j S i n G
i j
) )

K n
+ ^ (D cos 9 + C sin 9 )
M Q i*k k i k l k l k

- z^ (D k j cos e k j + sin e k j )

Let P = P - D,, - ^ Z (P . - S (D cos 9 + C s i n 9 )) (4.8)

k mk kk M i # k mi i j Ij l j i j

then

W = P
k ~ ^ b S l n
k " 3 C O S
\ ] ( 4
* 9 )

where

KL n n
a = _ S (D cos 9 + C., s i n 9 ) - Z (C sin 9 - D cos 9 )
M n i*k l k 1 k
j*k 1 J J
 60

n M, n
b = (D s i n 9 + C cos 0 ) - _J Z (D., s i n 9 - C,. cos 9 )
j*k k 3 2
M Q i*k l f c
i ik i

E q u a t i o n ( 4 . 9 ) can be w r i t t e n i n a more convenient form

\ \ - k- k p T s i n
<\ - V ( 4 a o )

where

- 1 a

a = tan _
k
b
. 2 _,_ 2 l / 2 v x
and T, = (a + b )
k

Equation (4.10) i s a simple form representing the motion of the

c r i t i c a l machine f o r a c e r t a i n d i s t u r b a n c e . S i n c e we assumed t h a t the r e s t

of the system i s not responding t o the d i s t u r b a n c e , i t i s r e a s o n a b l e t o use

the p r e - d i s t u r b a n c e a n g l e s q t o c a l c u l a t e t h e parameters P^, and c t ^ .

The s t a b l e and u n s t a b l e e q u i l i b r i u m p o i n t s of e q u a t i o n (4.10) can be

e a s i l y computed by s o l v i n g e q u a t i o n (4.11) f o r 9^

P
k " k
T S i n ( Q
k " V =
(4-1D

and the u n s t a b l e e q u i l i b r i u m p o i n t (UEP) I s

9 U
= n - ef (4.12)
k k

We n o t e h e r e t h a t we have two s e t s of the parameters P^, T r and a^; one s e t

f o r the f a u l t - o n network and another f o r the p o s t - f a u l t network.

 61

4.2.1 The T r a n s i e n t Stability Region

During the transient period, an exchange o f energy takes place

between the r o t o r of the c r i t i c a l machine and the p o s t - f a u l t network. The

kinetic energy generated by the a c c e l e r a t i n g power during the f a u l t - o n

p e r i o d must be f u l l y absorbed by the p o s t - f a u l t network i n o r d e r t o m a i n t a i n

stability.
C O

Multiplying equation (4.10) by and i n t e g r a t i n g between 0^ and

with respect t o time f o r the f a u l t - o n network parameters we o b t a i n the

k i n e t i c energy generated by the fault

2
k
- e =
\ \l
Q = p
k ( e
k k
e )
- k
T [ c o s ( e
k - 4) - c o s( e
k k ) ]

(4.13)
f f f C
where P^, T^ and oc^ are the system parameters f o r f a u l t - o n network and 0^ i s

the c l e a r i n g a n g l e .

The p o t e n t i a l energy o f the p o s t - f a u l t network i s d e r i v e d i n the same

manner by i n t e g r a t i n g e q u a t i o n (4.10) between and 0^ u s i n g the post-fault

network parameters, we o b t a i n

- IM^ = ?l (9j - e) - TP[cos(e^ - o^P) - cos(Q^ - o f ) ] (4.14)

Note that t h e speed d e v i a t i o n a t the u n s t a b l e e q u i l i b r i u m point i s zero

(e = o ) .
 62

Equation (4.14) represents the energy function of the c r i t i c a l machine

d u r i n g the t r a n s i e n t p e r i o d . The L.H.S. r e p r e s e n t s the f a u l t k i n e t i c energy

and the R.H.S. r e p r e s e n t s the p o t e n t i a l energy f o r the p o s t - f a u l t network.

The energy balance equation f o r c r i t i c a l c l e a r i n g becomes:

I M - pP e - T cos (Qp - o) + k U
= 0 .. (4.15)

where k U
= P 9^ + T cos (GJJ - c)
r
a g a i n we r e p r e s e n t 9^ by T a y l o r s e r i e s expansion

and

k
e =
Vc

where Y, =
k w
1 [P,
k
-
ek
P^Ct-.
OH
v ]

" 1 . 2

L e t X =
~ Vc
By r e p l a c i n g c o s ( 9 ^ - ot^) i n e q u a t i o n (4.15) w i t h the c o s i n e s e r i e s

expansion up to the f o u r t h o r d e r , we o b t a i n

M Y,x - pP(9 + ) X - T J[1 - 1 L_+ 1 L_ + k U

= 0
2! 4!

Let P - 9 - aP

then we get the c a t a s t r o p h e m a n i f o l d equation

 63

k
+ i p x 3
+ i ( i - 5l) x
2
+ (MY, - pP) + 2
Tl B- i f3 )x
3

24 6 2 2 k k k
6

+ ( k u
- T P - p P e + ! k
p 2
- ^ p 4
) = 0 (4.16)
2 24

24
M u l t i p l y e q u a t i o n ( 4 . 1 6 ) by - to give:
T P
f

2 T
X 4
- 4BX 3
- 12(1 - *L)X 2
- I 4
(MY, - pP + B- A p )X
3

2 T p k k k
6
k

(4.17)
~ ( k U
" k - k k
T P 9 +
~ P "~ 2
-0

Tp 2 24
A
k

We need t o e l i m i n a t e the t h i r d order term i n o r d e r t o have e q u a t i o n (4.17)

i n the s w a l l o w t a i l c a t a s t r o p h e m a n i f o l d form

Let A 3 = - 4B

A 2 = - 12(1
2

A x = - (MY - P j + TP B -
2
k B ) 3

4
 64

A Q - - 1*
T P
(k U
- - 0P +l l
T

2
B 2
T
- B )
4

24
X
k

and x = y - u (4.18)

3
A

u = _ (4.19)
4

We o b t a i n the s w a l l o w t a i l c a t a s t r o p h e manifold

4 2

y +uy + u y + w = 0 (4.20)

3 2
where u = (A - A )
8
. (A, - ^ I t f L )
1
2 8
2
/A A
1 3 A
. 2 3
A A
3 . 4.
w = (A - + - A- )
U
4 16 256

The b i f u r c a t i o n s e t B can then be d e f i n e d by

4y 3
+ 2uy + u = 0 ... (4.21)

The transient stability region is formed in the shape of the

swallowtail bifurcation set (see Appendix A) bounded by the transient

stability limits. The region i s defined i n terms of the system parameters

which can be e a s i l y calculated.

In summary, the procedure of calculating the transient stability

regions for multi-machine power systems using the critical-machines

dynamic-equivalent method i s o u t l i n e d i n the f o l l o w i n g steps:

 65

1. D e f i n e the c r i t i c a l machine(s) f o r the f a u l t considered.

2. C a l c u l a t e the parameters P , T fc fc and t o form the dynamic e q u i v a l e n t of

the c r i t i c a l machine(s).

3. Calculate the c a t a s t r o p h e m a n i f o l d parameters u, u and w from which the

critical clearing time and the degree of s t a b i l i t y are determined.

It i s important to note that the transient stability region using

this method is the same shape regardless of fault location and loading

condition. T h i s makes the method a good c a n d i d a t e f o r o n - l i n e assessment of

the transient s t a b i l i t y and the degree of s e c u r i t y of power systems.

4.3 T a y l o r E x p a n s i o n o f the A c c e l e r a t i n g Power

The chaotic nature of the swing equations of multi-machine power

systems i s a major d i f f i c u l t y i n s o l v i n g the t r a n s i e n t problem and i n t r y i n g

to i n c l u d e more d e t a i l e d system models In d i r e c t methods.

Recalling the swing equation of machine i i n an n-machine power

system from s e c t i o n (4.2):

V i - mi P
< ij
D C O S 6
i j + C
i j S l n 6
ij>

= P fli ( i = 1, ... n) (4.22)

where P ^ i s the a c c e l e r a t i n g power of machine i .

In the conventional time solution, equations (4.22) are solved by

numerical integration f o r the t r a n s i e n t p e r i o d (t n -* t ), from the i n s t a n t

 66

of fault o c c u r r e n c e t o the u n s t a b l e e q u i l i b r i u m point. I t i s already known

that each post-fault network has a f i x e d energy-absorbing capacity, called

the critical energy [51]. The c r i t i c a l energy can be c a l c u l a t e d off-line

for any p o s t - f a u l t condition. This energy i s the n u m e r i c a l i n t e g r a t i o n of

the post-fault accelerating power from the s t a b l e equilibrium point t o the

unstable equilibrium point.

We suggest u s i n g Taylor s e r i e s expansion t o f i n d the energy function

as a function of time by expanding the a c c e l e r a t i n g power e q u a t i o n around

the instant of f a u l t o c c u r r e n c e a t t=0. This p r o v i d e s the advantage of

eliminating the c a l c u l a t i o n s of the s t a b e l and u n s t a b l e e q u i l i b r i u m points

on-line and s i m p l i f i e s the procedure o f d e f i n i n g the c r i t i c a l machines.

The energy function of machine i in a multi-mchine power system

during the t r a n s i e n t p e r i o d i s g i v e n by:

t t t
c f c u
V
i =
f P
a l ( t ) d t +
f P
a i ( t ) d t +
' P
a i ( t ) d t ( 4
' 2 3 )

where P ^ i s fault-on accelerating power.

P f ( t ) i s post-fault accelerating power,

ai

t i s the c l e a r i n g time.

t ,t i s the times a t SEP and UEP.

s u

The first two terms of the R.H.S. of e q u a t i o n (4.23) a r e the k i n e t i c

energy generated d u r i n g fault-on, and the t h i r d term i s the c r i t i c a l energy

of the p o s t - f a u l t network. The p o r t i o n of the k i n e t i c energy needed t o move

 67

the r o t o r from 6 t o 6 of the p o s t - f a u l t network does not c o n t r i b u t e t o the

rt
0 s r

instability [51]. Therefore, i t should be subtracted from the energy

function i n o r d e r t o w r i t e the energy balance equation; this portion i s

t
s
/ P J ( t ) dt
n (4.24)

0 a l

Subtracting (4.24) from (4.23), and assuming that machine i I s the c r i t i c a l

machine and the f a u l t i s cleared at c r i t i c a l c l e a r i n g , we o b t a i n the energy

balance equation,
c t t
u
V. = / (pf(t) - pj(t))dt + / PJJ(t) d t = 0
1 a Qa t a
8
(4.25)

Equation (4.25) i s e v a l u a t e d by r e p l a c i n g P ( t ) and P ( t ) w i t h P

their
3. 3.
T a y l o r s e r i e s e x p r e s s i o n s , i n the form:

P (t)= P ( 0 )
+ i -P ( 1 )
t + i - P >t ( 2 2
+ (4.26)
a a u a 2, a

where P ( m )
= 1 (4.27)
a t=0
dt m

and the a c c e l e r a t i n g power of machine i

n
P - P - E (D . cos 6 + C.. s i n 6 )
ai mi i j i j i j i j

We note that a t t=0

= 0
 68

and 6^ = 6 ^ ( I n i t i a l o p e r a t i n g condition)

Therefore P , ( 0 )
= P - P . ( t = 0)
ai mi ei

n
= P (4.28)
mi j = \ < ij D c
s 6
i j + c
i j s i n

(1) _ d
= Z (D.. s i n 5 - C.. cos 6.1)6..
ai
dt 3 1

t -0 J-l i j i j i J i J i j

Since = 0
ij t-0

Therefore P f * = 0 1
(4.29)
ai

* - - 2i
t=0
^ (D j cos 6 + C j sin 6 ^ ) 6 ^

al d t ai

( sin 6 i
ij i j i j
+ D i j C C O S 6 ) 6

Since 6 = 0

n
n c 0 c 0
P ! 2 )
= S (D sin 6 (4.30)
ij i j i j

ai
C C S 6 6
j
J-l

where 6.^ = p ( i n i t i a l acceleration)

Ij al I

The derivations of the r e s t of the T a y l o r series coefficients are

g i v e n i n Appendix C. I t i s found that a l l odd o r d e r c o e f f i c i e n t s a r e z e r o s ;

 69

t h i s reduces the number of T a y l o r s e r i e s c o e f f i c i e n t s to one h a l f .

Therefore,

P
al = a<*
d i
P 0 ) + L p

2!
i 2 >
< 2 +
M
4!
4 )
' 4 +
- M-'D
The accelerating power c o e f f i c i e n t s have to be calculated f o r the

system d u r i n g the f a u l t and f o r the post f a u l t condition.

Usually i n large power systems the c r i t i c a l clearing time i s very

small (typically < .5 second). F o r t h i s p e r i o d we w i l l use up to the s i x t h

order of the a c c e l e r a t i n g power expansion f o r both f a u l t - o n and p o s t - f a u l t

conditions. Thus e q u a t i o n (4.25) becomes;

C
c D P ( 2 ) f
- P ( 2 ) P
P ( 4 ) f
- P ( 4 ) P

V. - J (<Pl 0 )
- P< 0 )
) + 0 ^ _ ) t 2
t 4

1
0
a
2! 4!

p (6) _ (6)P
f
p

+ (_f 1 ) t )dt + /
6
P ( t ) dt - 0
P
(4.32)
6! t a

The last term i s the c r i t i c a l energy of each s p e c i f i c post-fault condition.

The critical energy i s evaluated off-line f o r each p o s t - f a u l t condition by

the t r a p e z o i d a l r u l e as f o l l o w s
t
v c r - J U
P j ( t ) dt
s

V (1+1) = V (k) +1 (P ( t ) ( k + l ) - P ( t ) ( k ) ) - ( t ( k + l ) - t ( k ) )
A
 70

4.3.1 Transient S t a b i l i t y Regions Using Taylor Expansion

The Taylor series expansion of the accelerating power provides

f l e x i b l e choice over the seven elementary catastrophes of Table (3.1). The

choice of a s p e c i f i c catastrophe type depends on the highest order of the

Taylor series that w i l l give the best required r e s u l t s . I t i s needless to

add that the lower order catastrophes are easier to v i s u a l i z e .

We start with the resultant Taylor series expansion of equation

(4.32).

P
( 2
> p >
( 4
.
p (t) = p (o) + t z
+ - i _ t* + ...
a a
2! 4!

The rotor angle can also be represented by the form

6 = 6
I Oi
+I T
Y
i
t 2

which provides good agreement with the time solution for t < .5 second

[52].
* c 1 2
We let x =
i " 0i = - Y
i fc

2
substitute for t i n the accelerating power series

p (2) p (4)
P (x) = P ( 0 )
+ _!_ X + _!_ X 2
+ ... (4.33)
a a
Y 2
Y
i 6y z

The energy balance equation can be evaluated with respect to the

angle advances x

x <
c
V = J P (x)dx + V =0
i Q a
 71

P ( 3 ) p (4) 3

= P ( 0 )
x + x 2
+ _i_JL + . . . + V = 0 (4.34)
3
2Y 2 C r

18

The c h o i c e of the determinant order d e c i d e s the c a t a s t r o p h e manifold

type as f o l l o w s

X 2
= fold

V 3
X = cusp
4 = swallowtail
X

5 = butterfly

Here we c o n s i d e r the cusp c a t a s t r o p h e m a n i f o l d g i v e n i n Appendix B.

The cusp c a t a s t r o p h e m a n i f o l d from e q u a t i o n (4.24) i s g i v e n by

p<4> P < >
2

v = a
x + 3 a
x +p 2 ( 0 )
x +V = 0

18Y 2
" i
3 T T P ' 2 1 8 Y 2
P < 0 )
1 8 r2

= x J
+ _ x z
+ a
X + IV = 0 (4.35)

P (4> p (A) P ( 4 ) ^
a a a

= X 3
+ ^X 2
+ A^X + AQ = 0 (4.36)

In order t o put e q u a t i o n (4.36) i n the standard cusp m a n i f o l d we need t o

2
e l i m i n a t e the second order term (X ) .

We l e t X = y - B

A
P=

We g e t
 72

3 A
2 2
2 1 2
A A

y J
+ (A - ) y + (A + - -i_) = 0 (4.37)
1
3 27 3

or

y + u y + u = 0 (4.38)

where

Equation (4.38) i s i n the standard form of the cusp c a t a s t r o p h e .

The bifurcation set i s defined by equation (4.38) and the s e t of

degenerate c r i t i c a l p o i n t s .

3y 2
+ u - 0 (4.29)

The b i f u r c a t i o n s e t g i v e n i n Appendix B c o n s i s t s of a l l s t a b l e p o i n t s

in terms of the system parameters u and u.

4.4 I d e n t i f i c a t i o n o f the C r i t i c a l Machines

Both methods presented in Sections (4.2) and (4.3) rely on the

accurate identification of the critical machines for a specified

disturbance.

Correct identification has been a c h i e v e d by c a l c u l a t i n g the u n s t a b l e

e q u i l i b r i u m p o i n t s f o r a l l machines i n the power systems; the machine which

 73

has the h i g h e s t unstable equilibrium point is identified as the critical

machine [47], Although this method provides correct identification, i t s

main drawback i s the c a l c u l a t i o n of the UEP's, which i s time consuming, and

in some cases, wrong answers may be o b t a i n e d [ 4 8 ] .

Another method i s t o use the a c c e l e r a t i o n a t the i n s t a n t of fault

occurrence as a first identification and then calculate the critical

clearing time f o r each machine. The machine with the lowest critical

c l e a r i n g time i s the c r i t i c a l machine [ 4 5 ] .

For s m a l l and medium s i z e power systems, the c a l c u l a t i o n of UEP's i s

not difficult and usually the c o r r e c t answer i s achieved in a reasonable

time, but f o r l a r g e power systems the r i g h t answer i s not guaranteed.

In this thesis, the c r i t i c a l machines are i d e n t i f i e d as f o l l o w s :

I. C a l c u l a t e the i n i t i a l a c c e l e r a t i o n f o r each machine as f o l l o w s

\ " ^ t mi " e i
P P ( t
>
i M

where Pg^Ctg^ *" ^ s t i e e

^ e c t r
i c a
i power output during fault a t the

instant of f a u l t occurrence.

ii. The machines which have h i g h and positive initial accelerations are

i n j e c t i n g k i n e t i c energy to the system; t h e r e f o r e , they a l l contribute

to the system Instability. These machines are combined t o form a

c r i t i c a l group.

iii. The c r i t i c a l e n e r g i e s of the c r i t i c a l machines are a l s o added t o g e t h e r

to form the g l o b a l e n e r g y - a b s o r b i n g c a p a c i t y of the c r i t i c a l group.

 74

In most cases c r i t i c a l machines can be I d e n t i f i e d e a s i l y by the f i r s t

acceleration, especially when the f a u l t i s close t o one of the g e n e r a t o r

terminals. The procedure of combining a group of c r i t i c a l machines has to

be carried out only when a f a u l t o c c u r s a t non-generator buses or f a r from

g e n e r a t o r buses.

4.5 Numerical Examples

In this section two examples are presented to demonstrate the

validity and advantages of the application of catastrophe theory to

transient stability assessment of power Bystems. Transient stability

r e g i o n s i n terms of the system parameters a r e g i v e n f o r each example. Three

and seven machine power systems a r e used, three-phase short c i r c u i t s are

considered at d i f f e r e n t locations and a comparison between the p r e s e n t e d

methods and the s t e p - b y - s t e p time s o l u t i o n i s g i v e n f o r each example.

Each short-circuit case considered i s evaluated by the following

steps.

i. I d e n t i f y the c r i t i c a l machine or machines as e x p l a i n e d i n S e c t i o n 4.4.

ii. Dynamic e q u i v a l e n t method; c a l c u l a t e the dynamic equivalent parameters

of e q u a t i o n (4.10) f o r the c r i t i c a l machine(s) (as g i v e n i n Section

4.2). The b i f u r c a t i o n s e t parameters and c r i t i c a l clearing time a r e

then calculated.

iii. Taylor s e r i e s method; calculate the T a y l o r series coefficients of the

accelerating power f o r the c r i t i c a l machine (as g i v e n i n S e c t i o n 4.3).

The bifurcation s e t parameters and the c r i t i c a l clearing time a r e then

c a l c u l a t e d and the case i s l o c a t e d on the t r a n s i e n t stability region.

 75

4.5.1 The Three-Machine System

This system with nine buses, three machines and three loads, i s

widely referred t o i n the l i t e r a t u r e as the Western Systems Coordinating

Council (WSCC) t e s t system.

A single-line diagram of the system i s shown i n F i g . 4.1.

Transmission line parameters and loads are g i v e n i n per u n i t on a 100-MVA

base i n Table 4.1. Generator data and i n i t i a l operating c o n d i t i o n s are

g i v e n i n T a b l e 4.2.

Admittance (pu)
Bus No. G B

Generators
1 1-4 0 - 8.446
2 2-7 0 - 5.485
3 3-9 0 - 4.168

Transmission 4-5 1.365 -11.604

... L i n e s 4-6 1.942 -10.511
5-7 1.188 - 5.975
6-9 1.282 - 5.588
7-8 1.617 -13.698
8-9 1.155 - 9.784

Shunt Admittances
Load A 5-0 1.261 - 0.263
Load B 6-0 0.878 - 0.035
Loda C 8-0 0.969 - 0.160
4-0 0.167
7-0 0.227
9-0 0.283

T a b l e 4.1 Network parameters of the three-machine system

 77

Generator data Initial Conditions

Gen. No. H P E
d mo
(MW/MVA) (pu) (pu) (pu) (degree)

1 23.64 .0608 2.269 1.096 6.95

2 6.40 .1198 1.6 1.102 13.49

3 3.01 .1813 1.0 1.024 8.21

Table 4.2 Generator data and i n i t i a l c o n d i t i o n s of the three-machine system

Three-phase s h o r t c i r c u i t s are c o n s i d e r e d at d i f f e r e n t locations. The

transient stability i s evaluated for each fault by the step-by-step time

solution, dynamic equivalent and Taylor series methods. A comparison

between the three methods i s given i n Table 4.3 i n terms of the critical

clearing time. Both methods show very good agreement with the time

solution.

Fig. 4.2 shows the a c c u r a c y of the T a y l o r s e r i e s method u s i n g terms up

to the second order term for the angle during the fault-on period. Fig.

4.3 shows a comparison between the Taylor series approximation of the

accelerating power and the time solution for different fault locations

during the t r a n s i e n t p e r i o d .

The transient s t a b i l i t y region using the dynamic e q u i v a l e n t method i s

shown i n F i g . 4.4. A l l s t a b l e cases are shown i n s i d e the r e g i o n i n terms of

the system parameters. F i g . 4.5 show the transient s t a b i l i t y region using

the T a y l o r s e r i e s method. S t a b l e p o i n t s are marked i n s i d e the region.

250.0

0.0 0.1 0.2 0.3 0.4 0.0 0.1 0.2 0.3 0.4 0.5
time i n seconds

F i g . 4.2 A c c u r a c y of T a y l o r s e r i e s d u r i n g the f a u l t - o n p e r i o d
u s i n g o n l y the second o r d e r term.
time s o l u t i o n
T a y l o r s e r i e s method
C.C.T. c r i t i c a l c l e a r i n g time
 0.3

time i n seconds
- Afferent fault locations
F i e . 4.3 The a c c e l e r a t i n g power f o r d i f f e r e n t r a
time s o l u t i o n
Taylor series
Fault at C r i t i c a l C l e a r i n g Time [second)

Bus # Time Solution Dynamic Taylor Series

(step-by-step) e q u i v a l e n t method method

4 .29-.30 .27 .29

5 .22-.23 .20 .22

6 .51-.|2 .46 .49

7 .11-.12 .1 .11

8 .28-.29 .26 .28

9 .40-.41 .38 .41

T a b l e 4.3 C r i t i c a l c l e a r i n g time by the time s o l u t i o n and the

proposed methods f o r the 3-machine power system.
 81

F i g . 4.4 The t r a n s i e n t s t a b i l i t y r e g i o n of the 3-machine power

system u s i n g the dynamic e q u i v a l e n t method. S t a b l e cases
are marked ( x ) i n s i d e the r e g i o n .
F i g . 4.5 The t r a n s i e n t s t a b i l i t y r e g i o n of the 3-machine
system u s i n g T a y l o r s e r i e s method w i t h s t a b l e cases
marked (x) i n s i d e the r e g i o n .
83
 84

4.5.2 CIGRE 7-machine T e s t System

The CIGRE 225 KV test system i s shown i n F i g . 4.6. I t has 10 buses

and 13 lines. Buses 1 through 7 are generating buses while loads are

located at buses 2, 4, 6, 7, 8, 9 and 10. Generator, l o a d and t r a n s m i s s i o n

line data are g i v e n i n T a b l e 4.4. The base v a l u e s used are 225 KV and 100

MVA.

Three-phase faults are applied and the transient stability is

e v a l u a t e d f o r each f a u l t . The critical c l e a r i n g time i s c a l c u l a t e d f o r each

case u s i n g the t h r e e methods. T a b l e 4.5 g i v e s a comparison between the time

solution and the two methods presented in this chapter in terms of the

c r i t i c a l clearing time.

The transient stablity regions are shown i n Figures 4.7 and 4.8.

Both methods show very good agreement with the time solution plus well

defined transient stability r e g i o n s i n terms of the system parameters valid

for a l l l o a d i n g c o n d i t i o n s and fault locations.

4.6 D i s c u s s i o n of R e s u l t s

The results of the two examples presented i n the previous section

have shown the feasibility of catastrophe theory application to

multi-machine power systems. The accuracy of the methods i s very good when

compared w i t h the s t e p - b y - s t e p time s o l u t i o n which i s used as the benchmark

for accuracy. The methods proposed are f a s t and g i v e the c r i t i c a l clearing

time through a single computation instead of using the time solution

method repetitively and human i n t e r p r e t a t i o n to e v a l u a t e a s i m i l a r case.

 85

GENERATORS

Bus p X M E 6
base
p
m
(MVA) (%)( ) (MW s /rad) (MW) (p.u) ()

1 100 7.4 6.02 217 1.106 7.9

2 100 11.8 4.11 120 1.156 -0.2
3 100 6.2 7.59 256 1.098 6.5
4 100 4.9 9.54 300 1.110 3.9
5 100 7.4 6.02 230 1.118 7.0
6 100 7.1 6.77 160 1.039 3.6
7 100 8.7 5.68 174 1.054 7.9

LOADS

Bus p Q Bus P Q
(MW) (MVar) (MW) (MVar)

2 200 120 8 100 50

4 650 405 9 230 140
6 80 30 10 90 45
7 90 40

LINES

Bus R X uC/2
(ohm) (ohm) (MS)

1 - 3 5 24.5 200
1 - 4 5 24.5 100
2 - 3 22.8 62.6 200
2 - 10 8.3 32.3 300
3 - 4 6 39.5 300
3 - 9 5.8 28 200
4 - 5 2 10 200
4 - 6 3.8 10 1200
4 ~ 9 24.7 97 200
4 - 10 8.3 33 300
6 - 8 9-5 31.8 200
7 - 8 6 39.5 300
8 - 9 24.7 97 200

( ) These v a l u e s i n c l u d e the t r a n s f o r m e r ' s r e a c t a n c e s

and a r e expressed on a 100 MVA base.

T a b l e 4.4 Data f o r the CIGRE 7-machine system (taken from Ref. [20])
F a u l t at C r i t i c a l C l e a r i n g Time [second)

Rus # Time Solution Dynamic Taylor Series

(step-by-step) e q u i v a l e n t method method

1 .35-.36 .34 .36

2 .41-.42 .40 .37

3 .39-.40 .38 .39

4 .50-.51 .52 .48

5 .35-.36 .34 .35

6 .52-.53 .51 .51

7 .33-.34 .33 .33

8 .44-.45 .42 .45

T a b l e 4.5 C r i t i c a l c l e a r i n g time by the time s o l u t i o n and the

proposed methods f o r the 7-machine power system.
87
88
 89

The Importance of the t r a n s i e n t stability regions provided by the

bifurcation s e t of the c a t a s t r o p h e m a n i f o l d i s not only i n terms of speed

and accuracy. I t p r o v i d e s another important dimension t o the t r a n s i e n t

stability problem-the stability limits. The method checks f o r the

violations of the t r a n s i e n t stability limits of the p o s t - f a u l t network i n

terms of the system parameters. Of c o u r s e , i f the l i m i t s a r e exceeded the

system i s unstable. This important f e a t u r e i s not p o s s i b l e w i t h existing

direct methods, i . e . , the e x i s t i n g direct methods cannot g i v e any s o l u t i o n

when the s t a b i l i t y l i m i t s o f the p o s t - f a u l t network are v i o l a t e d .

The a c c u r a c y of the T a y l o r s e r i e s method becomes a problem when t i s

higher than .5 second [52]. To obtain good accuracy beyond this limit

higher order terms have to be included i n the computations. This will

c o m p l i c a t e the s t a b i l i t y r e g i o n and slow down the c a l c u l a t i o n procedure. In

p r a c t i c e , however, t h i s problem i s v e r y r a r e . F a u l t s i n l a r g e power systems

are usually tripped i n a few c y c l e s , typically .1 second. Interest is in

the d e l i v e r y of maximum power a t low c l e a r i n g time w i t h o u t r i s k i n g the power

system s e c u r i t y . T h i s p r a c t i c a l c o n s i d e r a t i o n can e a s i l y be handled by the

proposed methods w i t h o u t l o s s of a c c u r a c y .

The off-line calculations of the c r i t i c a l energy f o r each case

considered can be improved by c a l c u l a t i n g the c l o s e s t unstable equilibrium

point f o r the c r i t i c a l machine. This reduces the computation time and

simplifies the procedure of f i n d i n g the c r i t i c a l energy [ 4 5 ] .

 90

CHAPTER 5

INCLUSION OF DAMPING, FLUX DECAY AND EXCITATION RESPONSE

The most challenging problem i n the application of direct methods to

transient s t a b i l i t y of power systems i s the v a l i d i t y of the s i m p l i f i e d model

used to represent power systems.

In this Chapter a s i g n i f i c a n t improvement i n the modelling problem i s

presented. The new model proposed includes damping, f i e l d flux decay and

excitation response.

5.1 Limitation of the C l a s s i c a l Model

Power utilities are hesitant to accept the direct methods of

transient stability assessment mainly because they raise doubts about the

v a l i d i t y of the c l a s s i c a l model. In this model i t i s assumed that the flux

l i n k i n g the main f i e l d winding remains constant during the transient. This

may be true only i f the exciter does not respond during the f i r s t swing

period (1 second or less). Modern excitation systems can reach full

response within .1 second, so that the c l a s s i c a l model assumption i s not

valid for such exciters. In fact, during the l a s t decade trends i n the

design of power system components have resulted i n more reliance on f a s t -

response and high c e i l i n g voltage exciters [53].

Although fast-response exciters are more desirable and widely used i n

modern power systems, none of the existing direct methods of transient

stability assessment have considered the excitation e f f e c t . The reason i s

 91

t h a t Lyapunov's f u n c t i o n becomes so c o m p l i c a t e d t h a t the d i r e c t methods l o s e

the m e r i t of being f a s t methods.

I n c l a s s i c a l model assumptions, damping e f f e c t s are u s u a l l y n e g l e c t e d

although i t affects directly the accelerating power. In some cases the

damping power has a p p r e c i a b l e v a l u e and so a f f e c t s the f i r s t swing stability

[54].

5.2 Damping

In power systems t h e r e are sources of p o s i t i v e damping which tend to

damp out oscillations resulting from disturbances. T h i s damping i s due to

the characteristics of the mechanical system, generator and loads. The

mechanical system damping r e s u l t s from the i n c r e a s e i n s h a f t torque w i t h the

decrease i n speed. The per unit damping torque coefficient i s d e f i n e d as

the negative of the per u n i t change i n torque f o r each per u n i t change i n

speed [54].

The generator electrical damping is associated with the currents

t h a t are induced i n the r o t o r windings due to a d i s t u r b a n c e or oscillations.

The damping torque due to the f i e l d winding i s usually s m a l l because of the

relatively long time constant. The damper winding component of the damping

torque i s q u i t e a p p r e c i a b l e and i t u s u a l l y a f f e c t s a c c e l e r a t i n g power d u r i n g

disturbances.

Cylindrical rotor generators may develop a p p r e c i a b l e torque due to

the eddy currents which are induced by asynchronous operltion, such as

s l i p p i n g p o l e s or oscillations.
 92

Loads are a l s o a source of damping i n power systems. I n d u c t i o n and

synchronous motors w i t h t h e i r mechanical s h a f t loads develop damping torque

during disturbances.

The damping power (D6) i s u s u a l l y added to the i n e r t i a l power i n the

swing equation. The damping coefficient D includes the v a r i o u s damping

power components, both mechanical and e l e c t r i c a l . The damping coefficients

are usually i n the range of 1-3 pu. This represents mechanical damping,

generator and load damping. Larger values are also reported i n the

literature [2].

The swing e q u a t i o n of machine i In an n-machine power system i s g i v e n

by

M. 6 + D 6 = P - P . . ..
i i i mi ei i = l , . . . n (5.1)

where D i s the damping coefficient.

The a c c e l e r a t i n g power becomes

P
ai - mi " P P
e i" D
*i ( 5
' 2 )

The same p r o c e d u r e of S e c t i o n 4.3, i s used t o expand P ^ & i n t o a Taylor

series to get the energy function during the f a u l t - o n period. Recalling

equations (4.26) and (4.27)

P ( ) = P
4 t
( 0 )
+ ! - p ( 1 )
t + !-p( >t 2 2
+ ...

p (m) = ai
a
, m t=0
dt

we have
 93

P = P , - Z (D, . cos 6 + C. . s i n 6 ) - D& (5.3)

ai mi j ij ij ij ij i
= 1

We note that at t = 0 , 6^ = 0 .

Therefore

ai
( 0 )
= P_, " E (D*4 cos 6 +c
'ij "ij
. sin
4 6)
'ij'
- D(0) (5.4)
" j=l m i i J

(1) _ d P
ai
ai
dt t=0

11
0
- - E (-D , s i n 6,, + C,, cos 6 6 v
D)6f>
ij
4
ij -ij ij ij

= - D 6|) (5.5)

(0)
(0) _ ai
where 6 (5.6)
M.

(2) _ ^ a i
ai
dt' t=0

n 0. ,? 0,2
- = (
" iJ
D C
S
^J ' iJ
C S n
'i? ( 6
i >

^ (-D j sin 6
+ cos 6 )
6^ - 0*6^

- E (D sin 6 - C cos 6 ) 6 - D*6^ 0

(5.7)
 94

P ( 1 )

where 6 0
= (5.8)
M
i

Derivation of p
a ^ ^ an
& p
a i ^ a r e
8i v e n
i n
Appendix D. We note here

that a l l Taylor series coefficients exist when damping i s taken into

account.

P, = P $ > 0
+ P < > t 1
+ ! ? l ^ t 2
+ ... (5.8)
ai ai ai 2J

The energy e q u i l i b r i u m s u r f a c e i s then formed u s i n g e q u a t i o n (4.23)

V
i =
' Q
( P
a i " ai> P d t +
I P
a? d t + V
cr= 0
<'>
5 9

T h i s becomes

= A t 4
4
+ ^ t 3
+ A^ 2
+ A^t + A Q + k g + V r - 0 (5.10)

whe re A - i - (P f ( n )
- P P ( n )
)
n . a
n

t
c
K = / P dt -
P
constant
s J
Q ai

Here i s truncated a f t e r the 4 t h order term. We e l i m i n a t e the 3 r d order

term t o o b t a i n the standard s w a l l o w t a i l catastrophe manifold (Appendix A) as

given i n S e c t i o n 3.4.1 to get

X 4
+ uX 2
+ vX + w - 0 (5.11)

The transient s t a b i l i t y region i s defined through of the b i f u r c a t i o n s e t i n

'y

terms of the system parameters u, v and w which i n c l u d e s the damping term.

 95

The e f f e c t of damping on the a c c e l e r a t i n g power d u r i n g the t r a n s i e n t

period i s shown i n F i g . 5.1 f o r a s h o r t circuit on the power system of F i g .

3.5. Another example i s g i v e n i n F i g . 5.2 f o r a s h o r t circuit a t bus 7 of

the three-machine power system given i n F i g . 4.1. I t i s clear t h a t the

damping tends to reduce the f a u l t kinetic energy and r e s u l t s i n more s t a b l e

systems.

5.3 E x c i t a t i o n Response and F l u x Decay

When a f a u l t o c c u r s near a g e n e r a t o r bus, t r a n s i e n t currents i n the

armature circuit induce other c u r r e n t s i n the r o t o r circuit which carries

the field current. The flux linking the armature circuit will decay

according to the e f f e c t i v e time c o n s t a n t of the f i e l d circuit. This time

c o n s t a n t i s i n o r d e r of s e v e r a l seconds a t no l o a d , and one second o r h i g h e r

under load. The flux decay decreases the g e n e r a t o r i n t e r n a l voltage and

hence reduces the s t a b i l i t y limits.

Both steady-state and transient stability can be improved by

excitation control systems [55]. Fast speed of response and h i g h ceiling

voltage exciters can p a r t i c u l a r l y improve transient stability. With the

help of fast transient forcing of excitation and the boost of internal

machine f l u x , the output power of the g e n e r a t o r can be i n c r e a s e d d u r i n g the

transient period. This reduces the accelerating power and results in

improved transient performance. Fast exciters also ensure that subsequent

swings are s m a l l e r than the f i r s t swing. This i s important it or modern low

inertia generators and weakly damped power systems where transient

instability a f t e r the f i r s t swing may o c c u r .

 96
_.
0i8 ___ .

o.o H 1 1 1 1 1 1 ' 1 ^ 1 ' 1

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35

time in seconds
F i g . 5.1

0.8 T I

0.0-i 1 1 " 1 " 1 < 1 1

1 1 r J

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
time in seconds
F i g . 5.2 E f f e c t of damping i n c l u s i o n on the accelerating power during
the fault-on period for two test cases.
 97

The vector diagram for the transient state of a cylindrical rotor

g e n e r a t o r i s g i v e n i n F i g . 5.3. We assume t h a t the i n t e r n a l v o l t a g e lags

behind the q - a x i s by a constant angle <t>^ a l l the time.

The equations of motion of generator i i n an n-machine power system

with f i e l d f l u x decay and e x c i t a t i o n response are expressed as f o l l o w s

n
M
i 6
i= P
mi " j = \ < ij
D C 0 S 6
i j + C
i j 6 l n 6
iJ>

and
dE'
T ' jd o i< ^ " E
fA<
fdi~
E
%q i -

d i~
x
^ ) d l1*4
' di (5.12)

where

T^ ^ Q : d - a x i s t r a n s i e n t open c i r c u i t time constant

E
qi : v o
^- t a
8 e a
l 8 the q - a x i s r e l a t e d to E^ by angle
n
4>^.

E
fdi : e x c i t a t
*- o n
voltage a p p l i e d to f i e l d winding.

X,.,X' : d - a x i s synchronous and transient reactance,

di di

I,. : d-axis current,

di

The terms of the swing equation were p r e v i o u s l y d e f i n e d . From the vector

diagram of F i g . 5.3, and s i n c e <t>^ i s c o n s t a n t

E . = E cos <t>
qi i i

Equation (5.12) becomes

i fdi d E E

T
doi - r " \ - di - V cx d
i . ( 5 a 3 )

dt cos 4>

Equation (5.13) can be m o d i f i e d i n terms of the power e q u a t i o n and assuming

that 4>^ i s s m a l l we o b t a i n [56]:

 98

F i g . 5.3 The v e c t o r diagram of the generator i n t e r n a l q u a n t i t i e s

f o r the t r a n s i e n t s t a t e .
 99

T
doi - - - v ( x
di- di>
x
, V i g
J
E
J
c o s 6
i J
+ b
i J
E
J
s i n 6
i J
) ( 5
- 1 4 )

dt cos^ j=l J J
J J J

The first term of the R.H.S. of e q u a t i o n (5.14) r e p r e s e n t s the e x c i t e r

response w h i l e the second and t h i r d terms r e p r e s e n t the drop i n the i n t e r n a l

voltage due t o t h e f l u x d e c a y . I t i s c l e a r that i f E ^ ^ i s f a s t and h i g h

then the drop of i n t e r n a l voltage ( E ^ ) due t o t h e f l u x d e c a y c a n be

eliminated.

A typical excitation voltage-time response curve i s shown i n F i g .

5.4. OA i s the generator rated load field voltage before disturbance

occurs. The s t r a i g h t l i n e AC i s drawn such that the area ACD i s equal t o

the area ABD e n c l o s e d by the a c t u a l response. The e x c i t e r response i s g i v e n

by the r a t e of Increase or decrease of the e x c i t e r v o l t a g e , i . e . , the s l o p e

CE *"* AO
determined by i n F i g . 5.4. The time, i n seconds, f o r the e x c i t e r
OE

voltage to reach 95 percent of c e i l i n g voltage i s known as e x c i t e r voltage

response time [57]. An e x c i t e r with a voltage response time of 0.1 second

or less i s taken t o be a f a s t h i g h i n i t i a l response e x c i t e r .

The e x c i t e r v o l t a g e response i s g i v e n by
-t/T
E... = E + (E - E ) ( 1 - e ) e
(5.15)
rdi o c o

where E i s the i n i t i a l e x c i t e r voltage

o

E maximum c e i l i n g voltage
c

T e x c i t a t i o n system time constant

Substituting by E... i n equation (5.14) we get

rdi
Exciter
Voltage

time i n seconds

5.* The t J P i d -citation ,olta e-ti*e

g response
 101

, d E, E - (E -E )e

dt cos <t>.

n
" E
i " ( X
di" di X )
\ i s
il l
E C O S 6
i j + b
i j j
E S l n 6
i j } ( 5
* 1 6 )

The f u n c t i o n F ( t ) w i l l now be expanded i n a T a y l o r s e r i e s to evaluate

equation (5.16) f o r the t r a n s i e n t period (t < .5s). F ( t ) i s expanded

around t=0 as f o l l o w s

F(t)
- F (0) + F a> ^ +
2 )
t 2 + _ (5.17)
t=0 2!

where

,(n) _ d F ( t ) u

t=0
dt

iTherefore

E n
F ( 0 ) . _o_ _ i ( 0 ) _ ( x ^_ + b^EjBlnfi^)
cos <t>,

(5.18)

F (D . d F(t)
dt t=0
 102

-tli
1 <V o E ) e
d E
i n

T cos 4^ dt M
j=l J J J

+ b i j E j cos 6^)6^

at t = 0

ij dt T 1

doi

Thus

. (E -E ) (0)
F ( 1 )
= l (5.19)
\ C
*1
S T
doi

p(2)^ F (3) ^ (n)

f ^ evaluated i n the same way. Equation (5.16) i s

evaluated using (5.17), and we get

E. = E(o) + F ( 0 )
t + I + I (5.20)
1
2 6

where E(o) i s the i n i t i a l generator i n t e r n a l v o l t a g e at t=to-.

Equation (5.20) r e p r e s e n t s the generator i n t e r n a l voltage during the

transient period including exciter response and flux decay. Truncation of

E^ depends on the length of the t r a n s i e n t p e r i o d . As we e x p l a i n e d i n the

previous chapter the Taylor series method can give good results up to a

period of .5 seconds. In p r a c t i c e , faults are u s u a l l y c l e a r e d f a s t e r than

the .5 s period (a typical clearing time in large power ^ systems is .1

second).
 103

Equation (5.20) can be used w i t h the method p r e s e n t e d i n Section 4 . 3

to i n c l u d e e x c i t a t i o n response and f l u x decay i n the assessment of t r a n s i e n t

stability. The swing equation of the critical machine i i n an n-machine

system i s g i v e n by

V i = P
mi~ f x
E
i j E
<Sij c o s
6
ij + b
ij s i n
j>
6 ( 5 > 2 1 )

= P
ai
From e q u a t i o n (5.20) l e t the i n t e r n a l v o l t a g e of the c r i t i c a l machine be

(0) F ( 1
>t 2
F ( 2 )
t 3

E = E(o) + F t U ;
t + _ + _ (5.22)

1
2 6

Substitute (5.22) in ( 5 . 2 1 ) and expand P i n a T a y l o r s e r i e s as i n S e c t i o n

4.3 to o b t a i n

P ( 2 )
t 2

P 4 P ( > + P < > t 1

+
a
...
ai a a 21

where P ( n )
= *A

dt n t =

The a c c e l e r a t i n g power c o e f f i c i e n t s are e v a l u a t e d and g i v e n as follows

P a
0 )
- P m i ~ ^ E (0)E
I j (g j cos 6 t + b j sin 6 )
(5.23)

p(D = - I F <>E. (g., cos 6 + b., sin 6 ) (5.24)

a j = 1 3 i j i j i j IJ
 104

P > - a
2
- ^ F U> E j ( 8lJ cos 6 ^ + b J sin 6^)

+ ^ E (0)E j (g j sin 6 - b j cos 6 ^ ) ^ (5-25)

The derivation of these coefficients and P ^ \ 3

P^ ^
4
a r e g i v e n i n Appendix
3 3

E.

The energy e q u i l i b r i u m surface (catastrophe manifold) can now be

w r i t t e n down i n terms of the a c c e l e r a t i n g power c o e f f i c i e n t s as i n e q u a t i o n x

(5.9)

V
i = '0
( P
a i" a? P ) d t +
' Q
P
a? d t + V
cr = 0
< " >
5 26

Equation (5.26) i s e v a l u a t e d and truncated after the 4 t h order term to get

the swallowtail catastrophe manifold. This truncation i s valid f o r t < .5

second., Thus e q u a t i o n (5.26) becomes

V. = A . t 4
+ a t 2
+ A . t + A_ + K + V = 0 (5.27)
I 1 0 s cr
0
H I

where

A = (P f ( n )
- P P ( n )
)
n i ai ai
n!

t
c
K = / P dt = constant
P

s J
Q ao

V = critical energy
cr
 105

Again we eliminate the t h i r d order term of equation (5.27) t o get the

standard s w a l l o w t a i l catastrophe (Appendix A) i n the form

X 4
+ uX 2
+ vX + w = 0 (5.28)

The effect of f l u x decay and e x c i t e r response on the a c c e l e r a t i n g

power during the t r a n s i e n t p e r i o d i s shown i n F i g . 5.4 and F i g . 5.5 using

examples shown i n F i g . 3.5 and F i g . 4.1. I t i s apparent t h a t the area under

the a c c e l e r a t i n g power i s reduced. T h i s area represents the k i n e t i c energy

generated by the f a u l t . Therefore, the degree of s t a b i l i t y i s i n c r e a s e d and

the critical c l e a r i n g time i s also increased. This means that when fast

exciters are used the system can generate more power f o r the same c l e a r i n g

time.
0.8-

0.6-

0.4 c l a s s i c a l model
proposed method

0.2-

0.0 I ' 1
0.00 0.05 0.10 0.15 0.20 0.25
time i n seconds
Fig. 5.5

0.8-

0.6-
c l a s s i c a l model
proposed method

0.4-

0.2

0.0 1 I
0.00 0.05 0.10 0.15 0.20 0.25
time i n seconds
F i g . 5.6
The effect of flux decay and e x c i t a t i o n response on the accelerating
power during the transient period for two test cases.
 107

CHAPTER 6

CONCLUSIONS

In this thesis, the t r a n s i e n t energy f u n c t i o n was efficiently used

with catastrophe theory to d e f i n e comprehensive t r a n s i e n t s t a b i l i t y regions

for power systems. Explicit boundaries of the transient s t a b i l i t y regions

were identified by the b i f u r c a t i o n set of the catastrophe manifolds. The

method i s made p r a c t i c a l by u s i n g the concept of c r i t i c a l machines s i n c e f o r

a specified fault, only a few machines show s i g n i f i c a n t oscillations during

the transient period.

The thesis suggests two methods to solve the transient stability

problem of multi-machine power systems. In the first method, the critical

machine(s) i s i d e n t i f i e d and a two-machine dynamic e q u i v a l e n t f o r the power

system i s formed. The c r i t i c a l machine i s s i n g l e d out and a center of angle

group i s formed f o r the r e s t of the system. The energy balance e q u a t i o n is

derived from the equation of motion of the critical machine against the

center of angle. At critical c l e a r i n g , the energy balance equation forms

the e q u i l i b i r u m s u r f a c e of the c a t a s t r o p h e manifold from which the t r a n s i e n t

stability region i s derived by the b i f u r c a t i o n technique. This stability

region is valid for different loading conditions and fault locations. This

method has the advantage t h a t the critical energy i s c a l c u l a t e d d i r e c t l y by

using the c l o s e s t u n s t a b l e equilibrium point. However, t h i s method can only

be used w i t h the c l a s s i c a l model. *

In the second method, the c r i t i c a l machine i s i d e n t i f i e d and a Taylor

series expansion for its accelerating power for a specified fault is

 108

constructed. The energy balance e q u a t i o n i s then d e r i v e d by i n t e g r a t i n g the

a c c e l e r a t i n g power. The c r i t i c a l energy i s c a l c u l a t e d o f f - l i n e by n u m e r i c a l

integration f o r the f a u l t considered. A g a i n the energy balance e q u a t i o n a t

critical clearing r e p r e s e n t s the c a t a s t r o p h e m a n i f o l d from which the trans-

ient stability r e g i o n i s determined. T h i s r e g i o n i s bounded by the stabili-

ty l i m i t s and i s v a l i d for different l o a d i n g c o n d i t i o n s and f a u l t locations.

T h i s method i s l i m i t e d to a p e r i o d of .5 second f o r best a c c u r a c y . However,

this i s adequate f o r most large power systems since typical clearing times

are i n the range of a few c y c l e s or .1 second. Model Improvements were made

possible using this method. Excitation response, f l u x decay, and damping,

were included in the stability analysis by the Taylor series expansion

method.

The results o b t a i n e d by the proposed methods are i n good agreement

with those o b t a i n e d by the time s o l u t i o n method. A l a r g e number of s i m u l a -

tions are presently needed i n system planning i n order to determine the

critical clearing time or stability limits. The proposed methods give

directly the c r i t i c a l clearing time and stability l i m i t s w i t h good a c c u r a c y

and less computation. T h e r e f o r e , they can be used to g r e a t l y reduce the

large number of time simulations i n the system planning stage. In system

operations, the proposed methods provide fast solutions to the transient

stability problem with definite stability boundaries so that corrective

a c t i o n can be taken to prevent instability.

The major c o n t r i b u t i o n s of t h i s r e s e a r c h are the f o l l o w i n g :

1. For the f i r s t time, the c a t a s t r o p h e theory i s a p p l i e d to the t r a n s i e n t

stability problem. Comprehensive transient stability regions are

 109

calculated i n terms of the power system parameters. These r e g i o n s are

valid f o r any loading condition and fault location. A l l existing

direct methods, on the o t h e r hand, require new computation f o r any

chf.nge i n o p e r a t i n g c o n d i t i o n s . This i s an important c o n s i d e r a t i o n i n

implementing direct methods for real-time assessment of transient

s t a b i l i t y and system security.

Another unique advantage of the methods p r e s e n t e d i s the i n c l u s i o n of

the excitation response during the t r a n s i e n t period i n the stability

analysis. A l l existing direct methods are l i m i t e d t o the classical

model [47] which n e g l e c t s the e x c i t e r response. With the i n c l u s i o n of

damping, flux decay and e x c i t a t i o n response i n the p r e s e n t e d methods,

all factors that directly effect transient stability are taken into

account. T h i s r e s u l t makes the d i r e c t methods more r e a l i s t i c .

The methods presented predict any v i o l a t i o n of s t a b i l i t y limits and,

therefore transient instability. This result i s very important; i t

helps to ensure the s e c u r i t y of power systems f o r any disturbance

considered. I t also enables power system p l a n n e r s to d e s i g n proper

stability c o n t r o l s to p r e v e n t system instability.

E x i s t i n g d i r e c t methods e i t h e r n e g l e c t or approximate the e f f e c t of the

transfer conductances. However, i n some cases transfer conductances

can have an a p p r e c i a b l e effect on system performance [50] . In this

r e s e a r c h , the t r a n s f e r conductances are f u l l y r e p r e s e n t e d .

The application of c a t a s t r o p h e t h e o r y i s extended to the steady state

stability problem of cylindrical-rotor generators. The method

 110

presented (Chapter 3) is also adaptable to include the effect of

governors on steady s t a t e stability.

6. The achievements of this research have made the on-line security

assessment more f e a s i b l e than ever b e f o r e . The important requirements

for real-time applications are speed, accuracy and information. The

methods p r e s e n t e d need minimal c a l c u l a t i o n s to l o c a t e the system p a r a -

meters w i t h respect to the stability region. Good agreement w i t h the

time solution method were obtained (Chapter 4). The l o c a t i o n of the

system parameters, for each case, on the transient stability region

gives enough information on the degree of stability of the power

system.

The outcome of t h i s t h e s i s g i v e s new m o t i v a t i o n to the a p p l i c a t i o n of

fast direct methods to the transient stability problem of power systems.

New areas of research need to be explored i n order to reach the ultimate

goal of the d i r e c t methods which i s o n - l i n e assessment of t r a n s i e n t stabili-

ty. The future research should i n c l u d e the following:

1. The calculation of the critical energy of the post-fault network is

still time consuming and needs to be done o f f - l i n e . With some approxi-

mations, the dynamic e q u i v a l e n t method can give f a s t answers by calcu-

lating the closest unstable equilibrium points. However, a fast and

exact method is s t i l l needed i n order to complete the overall on-line

approach.

2. S t a b i l i t y controls such as f a s t v a l v i n g , braking r e s i s t o r s , s i n g l e pole

switchings, series capacitors and generator trippings are usually

applied in practice to restore transient stability of power systems.

 Ill

The inclusion of these c o n t r o l s i n the proposed direct methods i s of

great i n t e r e s t to power utilities.

This research considered only single disturbances. Multiple distur-

bances should a l s o be c o n s i d e r e d i n f u t u r e r e s e a r c h .

More can a l s o be done i n the area of steady state stability. The

presented method (Chapter 3) i s adaptable to include speed governor

control systems. This Is an interesting area to i n v e s t i g a t e and an

on-line steady state stability assessment approach can be developed

u s i n g the c a t a s t r o p h e theory.
 112

REFERENCES

[1] E.W. Kimbark, "Power System Stability", V o l . I , John Wiley & Sons
Inc., New York, 1948.

[2] P.M. Anderson and A.A. Fouad, "Power System C o n t r o l and S t a b i l i t y " ,
V o l . I , Iowa S t a t e U n i v e r s i t y P r e s s , Ames, Iowa, 1977.

[3] G.W. Stagg and A.H. E l - A b i a d , "Computer Methods i n Power System

A n a l y s i s " , McGraw-Hill, 1968.

[4] C. C o n c o r d i a , " R e p r e s e n t a t i o n of Loads", IEEE PES Winter Meeting,

paper 75 CH0970-4-PWR, 1975.

[5] IEEE Committee Report, "Dynamic Load Dynamics S i m u l a t i o n E f f e c t s and

D e t e r m i n a t i o n of Load C o n s t a n t s " , Paper T73-089-0, IEEE T r a n s . PAS,
V o l . 92, M a r c h / A p r i l 1973, pp. 600-609.

[6] F. John Meyer and K.Y. Lee, "Improved Dynamic Load Model f o r Power
System S t a b i l i t y S t u d i e s " , IEEE PES, Winter Meeting 1982, 82 WM
126-1.

[7] A.M. Sasson, "The Power System S t a b i l i t y Problems and i t s S o l u t i o n

Methods", IEEE PES Summer Meeting, Los A n g e l e s , C a l i f . , J u l y 1970.

[8] W.D. Humpage and B. S t o t t , " P r e d i c t o r - C o r r e c t o r Methods of Numerical

I n t e g r a t i o n i n D i g i t a l Computer A n a l y s i s of Power System T r a n s i e n t
S t a b i l i t y " , P r o c . I E E , V o l . 112, pp. 1557-1565, 1965.

[9] N.D. Rao and H.N. Rao, " S o l u t i o n of T r a n s i e n t S t a b i l i t y Problems

through the Number S e r i e s Approach", P r o c . I E E , V o l . I l l , pp.
775-788, 1964.

[10] D.L. Johnson, J.B. Ward, "The S o l u t i o n of Power System Stability

Problems by Means of D i g i t a l Computers", T r a n s . AIEE-PAS, V o l . 75,
pp. 1321-1329, 1957.

[11] S.N. T a l u k d a r , " I t e r a t i v e M u l t i - S t e p Methods f o r T r a n s i e n t S t a b i l i t y

S t u d i e s " , IEEE Paper No. 70TP 48-PWR, Winter Power Meeting, New York,
1970.

[12] N.D. Rao and H.N. Rao, "Phase-Plane Techniques f o r the S o l u t i o n of

Transient Stability Problems", P r o c . I E E , V o l . 110, 1963, pp.
1451-1461.

[13] K.N. S t a n t o n and S.N. T u l u k d a r , "New I n t e g r a t i o n 'Algorithms f o r

Transient S t a b i l i t y S t u d i e s " , IEEE T r a n s . , V o l . PAS-89, May/June
1970, pp. 785-991.

[14] H.W. Dommel and N. Sato, "Fast T r a n s i e n t S t a b i l i t y S o l u t i o n s " , IEEE

T r a n s . , V o l . PAS-91, J u l y / A u g u s t 1972, pp. 1643-1650.
 113

[15] K. P r a b h a s h a n k e r and W. J a n i s c h e w s y j , " D i g i t a l Simulationof

M u l t i - M a c h i n e Power Systems f o r S t a b i l i t y S t u d i e s " , IEEE T r a n s . , V o l .
PAS-87, J a n . 1968, pp. 73-80.

[16] L.H. M i c h a e l s , "The AC/Hybrid Power System S i m u l a t o r and i t s Role i n

System S e c u r i t y " , P r o c . of the Power I n d u s t r y Computer A p p l i c a t i o n s
Conference, Boston, Mass., May 1971, pp. 128-136.

[17] A.H. E l - A b i a d and K. Naguppun, " T r a n s i e n t Stability Regions of

M u l t i - M a c h i n e Power Systems", IEEE T r a n s . , V o l . PAS-85, pp. 169-179.

[18] J . L . Willems and J.C. Willems, "The A p p l i c a t i o n of Lyapunov Methods

to the Computations of T r a n s i e n t S t a b i l i t y Regions f o r M u l t i - M a c h i n e
Power Systems", IEEE T r a n s . PAS., v o l . 89, No. 5/6, pp. 795-801,
May/June 1970.

[19] F.S. Prabhakara and A.H. E l - A b i a d , "A S i m p l i f i e d D e t e r m i n a t i o n of

T r a n s i e n t S t a b i l i t y Regions f o r Lyapunov Methods", IEEE T r a n s . PAS,
Vol.-94, No. 2, M a r c h / A p r i l 1975, pp. 672-689.

[20] M.A. P a i , "Power System S t a b i l i t y " , V o l . 3, N o r t h H o l l a n d Publishing

Co.

[21] M. R i b b e n s - P a v e l l a and F . J . Evans, " D i r e c t Methods f o r S t u d y i n g

Dynamics of Large-Scale E l e c t r i c Power Systems - A Survey",
Automatica, V o l . 21, No. 1, pp. 1-21, 1985.

[22] T. Athay, R. Podmore and S. Vivmani, "A P r a c t i c a l Method f o r D i r e c t

A n a l y s i s of T r a n s i e n t S t a b i l i t y " , IEEE T r a n s . , PAS-98, M a r c h / A p r i l
1979, pp. 573-584.

[23] A.A. Fouad and S.E. S t a n t o n , " T r a n s i e n t S t a b i l i t y of M u l t i - M a c h i n e

Power System P a r t I and I I " , IEEE T r a n s . PAS-100, pp. 3808-3424, Aug.
1981.

[24] M. R i b b e n s - P a v e l l a , Th. Nan Cutsem, R. O h i f o u i and B. Toumi, "Energy

Type Lyapunov-Like D i r e c t C r i t e r i a f o r Rapid T r a n s i e n t Stability
A n a l y s i s " , P r o c e e d i n g s of I n t e r n a t i o n a l Symposium on Power Systems
S t a b i l i t y , Ames, Iowa, May 1985.

[25] A.A. Fouad, " A p p l i c a t i o n of T r a n s i e n t Energy F u n c t i o n s t o P r a c t i c a l

Power System Problems", IEEE S p e i c a l P u b l i c a t i o n on Rapid A n a l y s i s of
T r a n s i e n t S t a b i l i t y , Power Winter Meeting, 1987.

[26] A. V a l v e e r , e t a l . , " R e l a t i v e T r a n s i e n t S t a b i l i t y Power System Margin

Index", IEEE PES Summer Meeting and EHV/UHV Conference, Paper C73
474-4, Vancouver, Canada, J u l y 1973.

[27] C K . Pang, F.S. Prabhakara, A.H. E l - A b i a d and A . J . K o i v o , "Security

E v a l u a t i o n i n Power Systems U s i n g P a t t e r n R e c o g n i t i o n " , IEEE PES
Winter Meeting, New York, Feb. 1973.
 114

[28] K.S. Fu, " S e q u e n t i a l Method i n P a t t e r n Recognition and Machine

L e a r n i n g " , Academic P r e s s , 1968.

[29] W. M e i s e l , "Computer Oriented Approaches to P a t t e r n Recognition",

Academic P r e s s , 1972.

[30] K.S. Fu and J.M. Mendel, "Adaptive L e a r n i n g and P a t t e r n R e c o g n i t i o n " ,

Academic P r e s s , 1970.

[31] G. Hadley, " L i n e a r Programming", Addison Wesley, 1962.

[32] P.H. Mengert, " S o l u t i o n of L i n e a r I n e q u a l i t i e s " , IEEE Trans, on

Computers, V o l . C-19, pp. 124-131, 1970.

[33] C L . Gupta, " T r a n s i e n t S e c u r i t y Assessment of Power Systems by

P a t t e r n R e c o g n i t i o n and Lyapunov's D i r e c t Method", Ph.D. t h e s i s ,
Purdue U n i v e r s i t y , West L a f a y e t t e , I n d i a n a , 1976.

[34] H. Hakkimmashhadi, "Fast Transient Stability Assessment", Ph.D.

t h e s i s , Purdue U n i v e r s i t y , West L a f a y e t t e , I n d i a n a , Aug. 1982.

[35] S. Yamashira, T. K o i k e , A.H. E l - A b i o d , "Fast T r a n s i e n t S e c u r i t y

Assessment and Enhancement Using P a t t e r n R e c o g n i t i o n " , P r o c . of 8 t h
PSCC, Aug. 1984.

[36] T. Poston and I . Stewart, "Catastrophe Theory and I t s A p p l i c a t i o n s " ,

Pitman P u b l i s h i n g Co., London, 1979.

[37] A.A. Sallam and J . L . D i n e l e y , "Catastrophe theory as a Tool f o r

Determining Synchronous Power System Dynamic S t a b i l i t y " , IEEE T r a n s .
PAS, V o l . 102, pp. 622-630, March 1983.

[38] R. Thorn, " S t r u c t u r a l S t a b i l i t y and Morphogenesis", Benjamin-Addison

Wesley, New York, 1975.

[39] I . Stewart, "Elementary C a t a s t r o p h e Theory", IEEE Trans, on C i r c u i t s

and Systems, V o l . CAS-30, pp. 578-586, August 1983.

[40] P.T. Saunders, "An I n t r o d u c t i o n t o Catastrophe Theory", Cambridge

U n i v e r s i t y P r e s s , 1980.

[41] R.F. Chu, D i s c u s s i o n to Reference [ 3 7 ] , IEEE T r a n s . PAS, Vol.-102,

pp. 629-630, March 1983.

[42] M.D. Wvong and A.M. M i h i r i g , "Catastrophe Theory A p p l i e d t o T r a n s i e n t

S t a b i l i t y Assessment of Power Systems", Proc. I E E , V o l . 133, P t . C,
No. 6, Sept. 1986.

[43] M. R i b b e n s - P a v e l l a , Thn Van Cutsem, R. O d i f a o u i and B. Toumi, "Energy

Type L i a p u n i v - L i k e D i r e c t C r i t e r i a f o r Rapid T r a n s i e n t Stability
Analysis", P r o c of the I n t e r n a t i o n a l Symposium on Power System
S t a b i l i t y , Ames, Iowa, May 1985.
 115

[44] A.M. M i h i r i g and M.D. Wvong, " A p p l i c a t i o n of C a t a s t r o p h e Theory to

T r a n s i e n t S t a b i l i t y A n a l y s i s of Power Systems", P r o c e e d i n g s of LASTED
I n t e r n a t i o n a l Conference on High Technology i n the Power I n d u s t r y ,
Bosman, Montana, August 1986.

[45] Y. Xue, Th. Van. Cutsem and R i b b e n s - P a u e l l a , "A Simple D i r e c t Method

f o r Fast T r a n s i e n t S t a b i l i t y Assessment of Large Power Systems", IEEE
PES Winter Meeting, New O r l e a n s , L o u i s i a n a , February 1-6, 1987.

[46] R. Rahimi and G. S c h a f f e r , "Power System T r a n s i e n t S t a b i l i u t y Indexes

f o r On-Line A n a l y s i s of 'Worst Case' dynamic C o n t i n g e n c i e s " , IEEE PES
Summer Meeting, Paper 86 SM332-1, Mexico C i t y , Mexico, J u l y 20-25,
1986.

[47] A.N. M i c h e l , A.A. Fouad and V. V i t t a l , "Power System Transient

S t a b i l i t y U s i n g I n d i v i d u a l Machine Energy F u n c t i o n s " , IEEE T r a n s . ,
CAS-30, May 1983, pp. 266-276.

[48] N. Kakimoto, Y. Ohnogi, H. Matsuda and H. Shibuya, " T r a n s i e n t

S t a b i l i t y A n a l y s i s of L a r g e - S c a l e Power System by Lyapunov's D i r e c t
Method", IEEE/PES, 83SM 403-3, J u l y , 1983.

[49] A.A. Fouad, " A p p l i c a t i o n of T r a n s i e n t Energy f u n c t i o n s to P r a c t i c a l

Power System Problems", IEEE S p e c i a l P u b l i c a t i o n on Rapid A n a l y s i s of
T r a n s i e n t S t a b i l i t y , No. 87 TH0169-3-PWR.

[50] V.R. Sastry, "Validity of N e l g e c t i n g Transfer Conductances in

T r a n s i e n t S t a b i l i t y S t u d i e s " , P r o c IEE, V o l . 120, No. 12, December
197?,

[51] A.A. Fouad and S.E. S t a n t o n , " T r a n s i e n t S t a b i l i t y of A M u l t i - M a c h i n e

Power System", Part I and I I , IEEE T r a n s . , V o l . PAS-100, pp.
3408-3424, J u l y 1981.

[52] S. Furuya and S. Iwamoto, " F a s t T r a n s i e n t S t a b i l i t y S o l u t i o n Using

T a y l o r Expansion and Energy F u n c t i o n " , EE i n Japan, V o l . 105, No. 3,
1985.

[53] C. C o n c o r d i a and P.G. Brown, " E f f e c t s of Trends i n Large Steam

T u r b i n e Generator Parameters on Power System S t a b i l i t y " , IEEE T r a n s .
PAS-90, pp. 993-1000, 1971.

[54] S.B. C r a r y , "Power System Stability", V o l . 2, John Wiley & Sons,

Inc., New York, 1947.

[55] IEEE Work Group on S p e c i a l S t a b i l i t y C o n t r o l s PSEC, ?k B i b l i o g r a p h y

on the A p p l i c a t i o n of D i s c r e t e Supplementary C o n t r o l s to Improve
Power System S t a b i l i t y " , IEEE T r a n s . , V o l . PWRS-2, No. 2, May 1987.

[56] N. Kakimoto, Y. Ohsawa and M. H a y a s h i , " T r a n s i e n t S t a b i l i t y A n a l y s i s

of Multi-machine Power Systems w i t h F i e l d F l u x Decays V i a Lyapunov's
D i r e c t Method", IEEE T r a n s . , V o l . PAS-99, No. 5, Sept/Oct. 1980.
 116

[57] M.S. Sarma, "Synchronous Machines (Their theory, Stability and

Excitation Systems)", Gordon and Breach Science Publishers, New
York.
 117

APPENDIX A

The swallowtail catastrophe:

The potential function i s

V(X) = X + uX + vx + wX

The e q u i l i b r i u m s u r f a c e M i s the hyper surface

5X + 3uX + 2uX + w = 0 (A.l)

and the s i n g u l a r i t y s e t i the subset of M f o r which the e q u a t i o n

20X + 6ux + 2v = 0 (A.2)

The b i f u r c a t i o n s e t , B, i s a t h r e e - d i m e n s i o n a l surface i n the c o n t r o l

space u, u and w.

Since we are concerned only with the q u a l i t a t i v e behaviour of the

system, and t h e r e f o r e want p r i m a r i l y to be a b l e to p l o t the b i f u r c a t i o n s e t

B. Let C be a p l a n e u = constant i n C ( t h e c o n t r o l space), B w i l l be a

u u

surve i n C, and i f we can s k e t c h t h i s curve f o r a l l v a l u e s of u we can b u i l d

up the complete s u r f a c e B.

Equation (A.2) i m p l i e s that u i s an odd f u n c t i o n of X, and t h a t i s

together with (A.l), implies t h a t w i s an even f u n c t i o n X. Hence w i s an

even function o f u, and so f o r any u t h e curve B u i s symmetric about the

w-axis.

Nest we d i f f e r e n t i a t e ( A . l ) and (A.2) o b t a i n i n g

dw (A.3)
dX
 118

and = - (30X 2
+ 3u) (A.4)
dX

the rest of the term i n (A.3) having v a n i s h e d i n account of ( A . 2 ) . We now

have to c o n s i d e r the cases u > 0 and u < 0 s e p a r a t e l y .

If U i s positive, then cannot vanish. Hence u is a strictly

dX
monotone f u n c t i o n of X and the e q u a t i o n

^ = - 2 (A.5)
du

is valid everywhere. Moreover, equation (A.2) i m p l i e s that X u < 0 with

e q u a l i t y o n l y when X = u = 0, a t which p o i n t w a l s o v a n i s h e s .

It f o l l o w s that B y i s smooth, t h a t w i s l a r g e when |x| i s l a r g e , and

dw
that the s i g n a l of i s t h e same as t h a t of u, v a n i s h i n g o n l y at the
du
origin. T h i s enables us to draw F i g . ( A . l ) .

If u i s negative, then vanishes f o r two r e a l v a l u e s of X, /

dX 10
dw
Consequently, vanishes f o r three v a l u e s o f X, these two t o g e t h e r w i t h
dX

X = 0 as b e f o r e , and i t f o l l o w s t h a t B u has a c r i t i c a l p o i n t a t X = 0 and

cusps a t the o t h e r two p o i n t s .

To determine the type of the c r i t i c a l p o i n t , we n o t i c e t h a t E q u a t i o n

(A.2) i m p l i e s t h a t f o r |x| < / the product X u cannot be n e g a t i v e . Since

10

X and u a l s o v a n i s h t o g e t h e r , i t f o l l o w s t h a t i f u i s s m a l l and p o s i t i v e so

is X, and i s then negative. This together with the f a c t that B u is

du
 119

symmetric about the w-axis, establishes that the c r i t i c a l point is a

r e l a t i v e maximum.

Finally, we n o t e that i f u = 0 then e i t h e r X = 0 o r X = / -1^.. We

10

have just seen that X = 0 corresponds to a maximum a t the o r i g i n , and

substituting into equation (A.l) we find that both the other roots give

9u 2

w = . Hence B has a p o i n t of s e l f - i n t e r s e c t i o n on the p o s i t i v e w-axis.

20 u

We then check that |x| large implies t h a t both | u| and w a r e a l s o l a r g e and

then, u s i n g the values of the parameter X t o t e l l us t h a t the order of the

points we have found i s : s e l f - i n t e r s e c i t o n , cusp, maximum, cusp

self-intersection, we can draw F i g . (A.2). And s i n c e the e q u a t i o n of the

l i n e of p o i n t s of s e l f - I n t e r s e c t i o n i s the p a r a b o l a

v - * L2 . u 0
I
20
We can put the curves B u together t o form the s u r f a c e B shown i n F i g .

(A.3). The o r i g i n of the name " s w a l l o w t a i l " i s now apparent.

To find the form of the p o t e n t i a l i n each o f the three regions into

which B d i v i d e s C, i t i s s u f f i c i e n t to consider points f o r which u = 0 and

u < 0 then the s o l u t i o n of E q u a t i o n ( A . l ) i s

X 2
= (- 3u / ( 9 u 2
- 20w))
10

There a r e three cases:

2
9u
(a) w > Equation ( A . l ) has no r e a l roots and V has no critical
20
 120

Fig.(A.2)
Fig.(A.1)

Cross s e c t i o n of the b i f u r c a t i o n set for

u > 0 and u < 0

r>f the swallowtail

The b i f u r c a t i o n set of tne *
 121

points.

9u 2
2
0 < w < Because /(9u - 20w) i s r e a l and l e s s than the r e a l and
20
2
positive -
3 u , both solutions for X a r e r e a l and p o s i t i v e and V has
four c r i t i c a l points, two maxima and two minima.
2

w < 0 Both solutions for X are real b u t one is negative.

Consequently V has o n l y two critical points, one minimum and one

maximum.
 122

APPENDIX B

The cusp catastrophe:

The p o t e n t i a l f u n c t i o n i s

V(X) = X 4
+ uX 2
+ uX (B.l)

so the e q u i l i b r i u m s u r f a c e i s a t h r e e - d i m e n s i o n a l space i n x, u and u g i v e n

by

4X 3
+ 2uX + u = 0 (B.2)

and the s i n g u l a r i t y s e t i s the subset of the e q u i l i b r i u m s u r f a c e such that

the d e r i v a t i v e of (B.2) i s a l s o e q u a l t o z e r o . I t i s g i v e n by

12X 2
+ 2u = 0 (B.3)

We find the b i f u r c a t i o n s e t by e l i m i n a t i n g the s t a t e v a r i a b l e X from (B.2)

and (B.3), we o b t a i n

8U" + 2 7 u
3 2
- 0 (B.4)

Equation (B.4) i s the projection of the three-dimensional manifold of

equation (B.2) onto the c o n t r o l space (u-u). The cusp manifold and the

b i f u r c a t i o n s e t i s shown i n F i g . ( B . l ) .

Equation (B.2) has three real roots within the b i f u r c a t i o n set

r e g i o n , o r when

8u 3
+ 27 u 2
< 0

But when

8u 3
+ 27u 2
> 0
 123

Fig. B.l The cusp m a n i f o l d and i t s bifurcation set.

 124

there i s only one real root.

Figure (B.2) shows the bifurcation set of u-u plane i n which the

f u c t i o n s V(X) i s sketched for different v a l u e s of the parameters u and u.

,. .* Tne cusp p o t e n t i a l f - c t i o n VCX) a t d i f f e r e n t
v a l u e s of the c o n t r o l v a r i a b l e s .
 126

APPENDIX C

Taylor s e r i e s expansion of the a c c e l e r a t i n g power:

The accelerating power of machine I i n an n-machine power system i s

given by ( E q u a t i o n 4.22)

P = P , - Z (D. . cos 6 + c. . s i n 6 ) (CI)

ai mi - 1 i j i j i j i j '

We expand P ^ i n the neighborhood of t=0 to o b t a i n ,

m
P , t ) - P<> + 1- P ^ t + 1- P < > t 3 2
+ ... + - p l ( n )
t (C2)
3.x 3 - . 3 3 . 3
1! 2! m!
T

where

,(m) _ a

m
dt t-0

We note t h a t a t t=0, w (speed d e v i a t i o n ) = 0

and 6. = 6 (initial angle)

n
therefore (C.3)
P
a 0 )
- V - <1] " s 6
iJ +
1J 6 l n 6
i3>

0 0,
- Z (D sin 6 - c . cos 6 )
3
dt ^ t-0 J-l 3 J J J

Since u. . = 0
t=0

therefore P (1) _
= 0 (C4)

To s i m p l i f y the d e r i v a t i v e e q u a t i o n s we l e t
 127

n
0 0
A = Z ( D ^ cos 6 jLj + c j s i n 6^)
(C.5)

and

B = 2 (D l j sin 6 - c j cos 6 ^ ) (C.6)

the d e r i v a t i v e s of A and B w i t h r e s p e c t t o time a r e

dt 1 J
dt J

(1) * (2)
f o r convenience let 6^ = 6^ y
, 6^ = 6 ... and so on

therefore

P<> = P , - A
a mi

dt a j

P<3) . d_ (1) = d_ (1)

dt 3
dt

A 6.!^ + B 6. . ( 2)
= B c (2)
"IJ Ij " 6 ^' (C.7)

(3) . i . . (D (D CD
3
P P (2) 2 A 6 6 (2) _ B 6 + A 6 6 (2) + B 6 (3)
a . a i j i j i j i j Ij IJ
dt

(C8)
= 0
 128

p CD P (1)
where 6< > 3
- - = 0
M
i M
j

P< > =
a
4

d
1_ P<
t a
> 3
= B 6< >
i j
4
+A6< >6< >
i j i j
1 3

+ 2A 6<> - B 6 J >

2
1
1
6 < > + 2 A S ^

2 1
^ , ^ - 2B 6 ^ ) 6<>

2

" 3
- \ f \ ? - A
\ ^

P >
( 4
- B 6 /*> + 3 A 6 <2) 2
(C9)
a i j i j

P< > - * _ P< > - B 6 / 5 ) + 5A 6 / / > 6 / / >

5 4
+
1 0 A
^ "> B 6//> 6//> 2

a d a i j i j i j i j i j i j i j

- 15 B 6 < > 6 < > - 10 A 6 < > \ < >1 j

2 2
i j
1
1
1 2
+ B 6 ^ - 0
(C.10)

P<> =
a d
1- P<
a
> 5
= B 6< >
i j
6
+ 6 A 6//>6//>+ 15 A 6 < > 6 <*>- 15 B
i j i j i j i j i j
2
6/^V/* i j

+ 10 A 6 < > - 60 B *
3
* h - 20 A 6 ^ 6 < > - 15 B 6 < >"
3
4
2

45 A 6 < >6 5 >

1
l
1
2 2
+ 15 B 6 " * + A 6 '

B 6 ^ + 15 A ^ 6 5 > 2)

4
- 15 B 6 < >
2 3
( C l l )
 129

Note that 6 ^ ) = 6^ > -

3
6 \

5 )
= 6 < > ... = 0

7
 130

APPENDIX D

D e r i v a t i o n of p
a ^ ^ a <* n p
a ^ ^ including damping:

P < ) - _3
<> P
2

ai ai d t

From e q u a t i o n (C.8) of Appendix C

(3)
P } ' (without damping) = 0
a l

Therefore

, - -> CD
Also

p {> - i - P >
al d t ai

From e q u a t i o n (C.4) we o b t a i n
P < > - B 6 <*>
4
+ 3 A 6 C2> - D 6 (5) 2

ai i j i j i j

Since 6 J
1
5 )
= 0

then

P <>
ai
4
- B 6 <*>
i j
+ 3 A a/
i j
)' 2

- ^ ( D ^ - m 6 - c
i j C os * ])*K
3

, 0. < ( 2 ) '
+ c l j S in A
6 ) j 6 J

(D.2)
 131

APPENDIX E

Taylor series coefficients of the a c c e l e r a t i n g power i n c l u d i n g excitation

response.

The accelerating power of machine i i s

P
a i - mi " ^ P
V j <lj C O S 5
ij + b
ij S i n 6
ij> ( E a )

and

E = E(0) + F ( 0 )
t + l i - i t + ... 2
(E.2)
1
2

Recalling equation (5.16)

d E
i E
c ~ ( E
c " V " t / T e

T
d 0 ; - - ^ F
-- e
- r<v- di>
E x

i dt cos 4>

S Ej ( g J cos 6 J + b j sin t 6 )
j (E.3)

the coefficients F ( 0 )
, F ( 1 )
are given i n Section (5.3), F ( 2
\ F ( 3 )
are

derived from (E.3) and g i v e n as follows.

F (2) = J__c
(E - E )
0 J (1)
_ + _ s
n
s i n 5 0
0

x 2

e
" S
* i W ' J = 1 j j j

b cs 6, CE.4)
 132

,(2)
,(3) _ 1 ( E
c " V (E.5)
z 3 cos
"doi
e

The d e r i v a t i o n of the a c c e l e r a t i n g power c o e f f i c i e n t s are as follows

Let A i = E j ( g ^ cos 6 ^ + b j s i n 6 )

(B.6)
B
ij j E (8
U 8 i n 6
ij " ij b C S
\^
d A
then M--B 6
dt
ij ij
d B
0

dt

I6 0 = 6 (1) g (1) = 6 (2)

Let
ij \i ' ij 6 6
ij ' * etc.

Substituting equations ( E . 6 ) , and (E.2) to ( E . l ) to o b t a i n the T a y l o r series

c o e f f i c i e n t s of P ^ we get

P<> - , , - I E<> A , , (E.7)

a mi i i j

,(D _ d P
ai = - S ( F ^ A . J, " E(0) B 2 61 J
dt t=0 j=l

i F ( 0
> A (E.8)
j=l ij
 133

>(2) _ d P
ai n
E (E(0) B 4 4 6^ 2 )
- F ( 1 )
A,,) (B.9)
dt t=0 j=l ij i j ij'

,(3) . ^ a i n
(E.10)
dt" t=0

4
ti\ d p
<

^ I 6f ( 1 > B
ij 6
i f } + 3 E
<> i j A 6
i f > 2

dt t=0

+ E(O) B j 6 <

4 )
- F ( 3 )
A j ] (E.ll)