HYGOV

Dynamic Simulations in Realistic-Size Networks
Pedro Rafael Barros Martins Arajo
Dissertao para obteno do Grau de Mestre em
Engenharia Electrotcnica e de Computadores
Jri
Presidente:
Prof. Doutor Paulo Jos da Costa Branco
Orientador:
Prof. Doutor Pedro Alexandre Flores Correia
Vogal:
Prof. Doutor Jos Manuel Dias Ferreira de Jesus
Outubro de 2010
Agradecimentos
Agradeo ao Professor Doutor Pedro Flores Correia, por todo o apoio, confiana e motivao
que me deu ao longo destes meses. Os seus conselhos foram um forte contributo para a
realizao deste trabalho, assim como o sero para a minha vida profissional.
minha famlia, agradeo a ajuda e o incentivo que me deu ao longo de todo o meu percurso
acadmico. Um especial obrigado aos meus pais, que sempre me apoiaram de forma
incondicional.
Por fim, deixo uma palavra de apreo a todos os meus amigos que, directa ou indirectamente,
me ajudaram a concretizar esta dissertao.
ii
Abstract
This dissertation is focused on the numerical implementation of dynamic simulations of electric
power systems for transient stability analysis.
Simulation and analysis of power systems is a crucial activity for power systems engineers,
which has become increasingly complex given the size of large interconnected networks, and
also given the demands in terms of security and quality of service.
The work developed for this dissertation is part of a continued effort in the development of a
student-grade program for transient stability analysis, implemented in MATLAB and reported in
previous dissertations.
The goal of this dissertation was to make the program capable of dealing with networks of
realistic sizes. Two main tasks emerged while pursuing that goal. Firstly, there was the need to
review the whole structure of the code in the program. Secondly, there was the need to revise
all the dynamic models developed so far in the program (the Round Rotor and Salient Pole
Synchronous Generators, the excitation controller "IEEE Type 1" and two governor systems,
one based on a steam turbine and another in a hydraulic turbine) and also the need to add new
dynamic models representing common components found in most networks, notably the speed
regulation system based on a gas turbine-governor and the excitation control system Type
DC1A. A complete revisitation of the numerical procedures for transient stability is also
performed. Finally, several simulations are carried out and the results that validate the models
and the ability of the program to deal with realistic size networks are presented. The validation
is performed with a side by side comparison with PSS/ETM.
Keywords: Power system analysis, Transient stability, Dynamic models, Exciter system,
Turbine-Governor systems.
iii
iv
Resumo
Esta dissertao centra-se na implementao de um programa de simulaes dinmicas de
sistemas energia elctrica para estudos de estabilidade transitria.
A simulao e anlise de sistemas de energia, que se tm tornado cada vez mais complexos
devido grande dimenso das redes interligadas e s exigncias de segurana e qualidade de
servio requeridas, uma actividade fundamental para os engenheiros de sistemas de energia.
O trabalho desenvolvido nesta dissertao est enquadrado num esforo contnuo para o
desenvolvimento de um programa acadmico dedicado ao estudo de estabilidade transitria,
implementado em MATLAB e descrito em dissertaes anteriores.
O objectivo desta dissertao foi o de tornar o programa capaz de lidar com redes de dimenso
realista. A concretizao deste objectivo requereu duas grandes intervenes. A primeira
consistiu na reviso da estrutura do programa. A segunda relacionou-se com a reviso dos
modelos dinmicos anteriormente desenvolvidos (o gerador sncrono de rotor cilndrico e o
gerador sncrono de plos salientes, o sistema de controlo de excitao "IEEE Type 1" e dois
reguladores de velocidade, um baseado numa turbina a vapor e outro numa turbina hidrulica).
Alm disso, adicionaram-se novos modelos dinmicos que representam componentes
frequentes nas redes elctricas, nomeadamente o sistema de regulao de velocidade
baseado num conjunto turbina-regulador a gs e o sistema de controlo de excitao Type
DC1A. Os procedimentos numricos para a estabilidade transitria foram tambm analisados.
Por fim, efectuaram-se vrias simulaes e os resultados que validaram os modelos e a
capacidade do programa em lidar com redes de dimenses realistas foram apresentados. Esta
apresentao feita lado a lado com os resultados do PSS/ETM, permitindo uma comparao
entre eles.
Palavras Chave: Anlise de sistemas de energia, Estabilidade transitria, Modelos dinmicos,

Sistema de excitao, Turbina-regulador.
vi
Table of Contents
Abstract ................................................................................................................................... iii
Resumo..................................................................................................................................... v
Table of Contents .................................................................................................................... vii
1.
2.
Introduction ...................................................................................................................... 1
1.1.
Context and Motivation ............................................................................................. 2
1.2.
State of the Art .......................................................................................................... 3
1.3.
Objectives.................................................................................................................. 5
1.4.
Dissertation Outline ................................................................................................... 6
Power System Simulation Software ................................................................................... 9

2.1.
2.1.1.
Simulation Data Acquisition ............................................................................. 10
2.1.2.
Power Flow Computation................................................................................. 10
2.1.3.
Modelling of the Network Equations ................................................................ 12
2.1.4.
Computation of the Initial Conditions for Dynamic Simulations ........................ 12
2.1.5.
Computation of the Algebraic Dynamic State Equations ................................... 13
2.1.6.
Dynamic Simulation ......................................................................................... 13
2.2.
3.
Simulation Software Improvements and Modifications ............................................ 15
2.2.1.
Reading of the Dynamic Data Files ................................................................... 15
2.2.2.
Presentation of the Results .............................................................................. 17
2.2.3.
Generator Reactive Power Limits ..................................................................... 18
Dynamic Models Description ........................................................................................... 21

3.1.
Round Rotor Synchronous Generator ...................................................................... 22
3.2.
Excitation Control Systems....................................................................................... 23
3.2.1.
Excitation Control System IEEE Type I, IEEET1................................................... 25
3.2.2.
Excitation Control System Type DC1A, IEEEX1 .................................................. 26
3.3.
4.
Simulation Software Algorithm ................................................................................ 10
Governor Control Systems ....................................................................................... 27
3.3.1.
Hydro-Turbine Governor, HYGOV ..................................................................... 27
3.3.2.
Gas-Turbine Governor, GAST............................................................................ 29
Differential-Algebraic Model ........................................................................................... 33

4.1.
Modified Euler-Cauchy integration algorithm .......................................................... 34
4.2.
Models in Differential Form ..................................................................................... 35
4.2.1.
GENROU and GENROE model ........................................................................... 35

vii
4.2.2.
IEEET1 model ................................................................................................... 36
4.2.3.
IEEEX1 model ................................................................................................... 37
4.2.4.
HYGOV model .................................................................................................. 38
4.2.5.
GAST model ..................................................................................................... 39
4.3.
4.3.1.
GENROU and GENROE models ......................................................................... 41
4.3.2.
IEEET1 model ................................................................................................... 42
4.3.3.
IEEEX1 model ................................................................................................... 43
4.3.4.
HYGOV model .................................................................................................. 44
4.3.5.
GAST model ..................................................................................................... 45
4.4.
5.
Initial Conditions for Dynamic Simulation ................................................................ 46
4.4.1.
GENROU and GENROE Initial Conditions........................................................... 46
4.4.2.
IEEET1 Initial Conditions ................................................................................... 48
4.4.3.
IEEEX1 Initial Conditions................................................................................... 48
4.4.4.
HYGOV Initial Conditions .................................................................................. 49
4.4.5.
GAST Initial Conditions ..................................................................................... 50
Simulation Results ........................................................................................................... 51

5.1.
Dynamic Simulation Procedure ................................................................................ 52
5.2.
Validation of the Dynamic Models ........................................................................... 52
5.2.1.
GAST Validation ............................................................................................... 54
5.2.2.
IEEEX1 Validation ............................................................................................. 56
5.2.3.
IEEET1 Validation ............................................................................................. 59
5.2.4.
HYGOV Validation ............................................................................................ 60
5.3.
6.
Models in Algebraic Form ........................................................................................ 40
57-Bus Case ............................................................................................................. 64
Conclusions ..................................................................................................................... 73
Summary of the Work ......................................................................................................... 73
Future Work........................................................................................................................ 75
Bibliography............................................................................................................................ 77
Appendices ............................................................................................................................. 79
Appendix A - Windup and Non Windup limits.......................................................................... 80
Appendix B - Exciter Saturation Function................................................................................. 82
Appendix C - Dynamic Parameters and Initial Conditions for Dynamic Simulations .................. 83
C.1.
GAST Validation ....................................................................................................... 83
C.1.1.
Dynamic Parameters ........................................................................................ 83

viii
C.1.2.
C.2.
Initial Conditions .............................................................................................. 83
IEEEX1 Validation ..................................................................................................... 84
C.2.1.
C.2.2.
C.3.
IEEET1 Validation ..................................................................................................... 84
C.3.1.
C.3.2.
C.4.
HGOV1 Validation .................................................................................................... 85
C.4.1.
C.4.2.
C.5.
57-Bus Network Simulation ..................................................................................... 86
C.5.1.
C.5.2.
Appendix D Algebraic State Equations for Validation Simulations ......................................... 89

D.1.
GAST Validation, GENROU + IEEET1 + GAST.............................................................. 89
D.2.
IEEEX1 Validation, GENROE+ IEEEX1 + TGOV1 .......................................................... 92
D.3.
IEEET1 Validation, GENROE+IEEET1+TGOV1 ............................................................. 95
D.4.
HYGOV Validation, GENSAL + IEEET1 + HYGOV ......................................................... 99
Appendix E 57-Bus Simulation Errors .................................................................................. 103

Appendix F 30-Bus Case ..................................................................................................... 111
F.1.
Dynamic Parameters.............................................................................................. 111
F.2.
Initial Conditions.................................................................................................... 112
F.3.
Simulation Results ................................................................................................. 113
ix
List of Tables
Table 1.1 Revisited dynamic models ..................................................................................................... 6
Table 1.2 Newly implemented models ................................................................................................... 6
Table 2.1 Array distribution ................................................................................................................. 16
Table 2.2 Column distribution ............................................................................................................. 16
Table 3.1 Parameters of the HYGOV dynamic model .......................................................................... 28
Table 3.2 Parameters of the GAST model ........................................................................................... 30
Table 5.1 Power flow results for the 2-Bus simulation .......................................................................... 53
Table 5.2 Power flow results for the 57-Bus simulation. ....................................................................... 65
Table 5.3 Dynamic models used in the 57-Bus simulation.................................................................... 66
List of Figures
Figure 1.1 Time ranges of dynamic phenomena (Source: [8]) ..................................................................... 4
Figure 2.1 Simulation process flowchart ................................................................................................... 11
Figure 2.2 Reactive Power limit verification function ................................................................................. 18
Figure 3.1 - Electromagnetic model of the round rotor generator, GENROU and GENROE (Source: [15]) ..... 23
Figure 3.2 Synchronous machines excitation subsystems (Source: [6]) .................................................... 24
Figure 3.3 Block diagram of the IEEET1 dynamic model (Source: [13]) ..................................................... 25
Figure 3.4 Block diagram of the IEEEX1 dynamic model (Source: [13]) ..................................................... 26
Figure 3.5 Speed Governor Basic scheme (Adapted from [9]) ................................................................... 27
Figure 3.6 Block diagram of the HYGOV dynamic model (adapted from [17]) ............................................ 28
Figure 3.7 Open system gas turbine ........................................................................................................ 29
Figure 3.8 - Block diagram of the GAST dynamic model (Adapted from [17]) ............................................... 30
Figure 5.1 2-Bus network ........................................................................................................................ 53
Figure 5.2 GAST model validation results (Part 1 of 2) ............................................................................. 54
Figure 5.3 GAST model validation results (Part 2 of 2) ............................................................................. 55
Figure 5.4 IEEEX1 model validation results (Part 1 of 2) ........................................................................... 57
Figure 5.5 IEEEX1 model validation results (Part 2 of 2) ........................................................................... 58
Figure 5.6 - IEEET1 model validation results .............................................................................................. 59
Figure 5.7 - HYGOV model validation results .............................................................................................. 61
Figure 5.8 Desired gate position state variable response ......................................................................... 62
Figure 5.9 Single-line diagram of the 57-Bus network ............................................................................... 64
Figure 5.10 57-Bus simulation results Voltages ..................................................................................... 67
Figure 5.11 57-Bus simulation results - Gen. group at bus 1, GENSAL + IEEET1 + HYGOV...................... 68
Figure 5.12 57-Bus simulation results - Gen. group at bus 2, GENROE + IEEET1 + GAST........................ 68
Figure 5.13 57-Bus simulation results - Gen. group at bus 3, GENROE + IEEEX1 + TGOV1 .................... 69
Figure 5.14 57-Bus simulation results Gen. group at bus 6, GENSAE + IEEET1 + HYGOV .................... 69
Figure 5.15 57-Bus simulation results Gen. group at bus 8, GENROU + IEEET1 + TGOV1 .................... 70
Figure 5.16 57-Bus simulation results Gen. group at bus 9, GENROU + IEEET1 + GAST ....................... 70
Figure 5.17 - 57-Bus simulation results Gen. group at bus 12, GENSAL + IEEET1 + HYGOV.................... 71
Figure A.1 Single time constant block - Windup limiter. ............................................................................. 80
Figure A.2 Single time constant block Non-Windup limiter...................................................................... 80
xi
Figure B.1 Saturation Function (Source: [12]) ........................................................................................... 82

Figure B.2 - Exciter saturation characteristics (Source: [5]) .......................................................................... 82
Figure E.1 57-Bus simulation error percentages Voltages .................................................................... 103
Figure E.2 57-Bus simulation error percentages - Gen. group at bus 1, GENSAL + IEEET1 + HYGOV .... 104
Figure E.3 57-Bus simulation error percentages - Gen. group at bus 2, GENROE + IEEET1 + GAST ...... 105
Figure E.4 57-Bus simulation error percentages - Gen. group at bus 3, GENROE + IEEEX1 + TGOV1 .... 106
Figure E.5 57-Bus simulation error percentages - Gen. group at bus 6, GENSAE + IEEET1 + HYGOV ..... 107
Figure E.6 57-Bus simulation error percentages - Gen. group at bus 8, GENROU + IEEET1 + TGOV1 .... 108
Figure E.7 57-Bus simulation error percentages - Gen. group at bus 9, GENROU + IEEET1 + GAST ...... 109
Figure E.8 57-Bus simulation error percentages - Gen. group at bus 12, GENSAL + IEEET1 + HYGOV .. 110
Figure F.1 30-Bus simulation results Voltage magnitudes .................................................................... 113
Figure F.2 30-Bus simulation results Gen. group at bus 1, GENSAL + IEEET1 + HYGOV .................... 113
Figure F.3 30-Bus simulation results Gen. group at bus 2, GENROE + IEEET1 + GAST....................... 114
Figure F.4 30-Bus simulation results Gen. group at bus 5, GENROE + IEEEX1 + TGOV1 .................... 114
Figure F.5 30-Bus simulation results Gen. group at bus 8, GENROU + IEEET1 + TGOV1 .................... 115
Figure F.6 30-Bus simulation results Gen. group at bus 11, GENSAE + IEEET1 + HYGOV .................. 115
Figure F.7 30-Bus simulation results Gen. group at bus 13, GENSAL + IEEET1 + HYGOV................... 116
xii
Nomenclature
Extension of the power flow input data
Extension of the dynamic input data files
Active Power
Reactive Power
Bus Voltage
Admittance matrix
Generators terminal current
Subscript
Denotes a bus injected quantity
Subscript
Denotes a reference to a generated variable

Reduced admittance matrix, provided by the internal node mode
Subtransient internal voltage
Machines complex power
d- and q-axis transient open-circuit time constant
d- and q-axis subtransient open-circuit time constant
or
d-axis transient reactance
or
q-axis transient reactance
or
d-axis subtransient reactance
or
q-axis subtransient reactance
or
and
Generators self reactance

Direct-axis damper winding
Quadrature-axis damper winding
Direct and Quadrature axis stator current
Direct and Quadrature axis stator voltage
Generator rotor angle
Nominal frequency
Generator rotor speed
Speed deviation
Mechanical torque applied to the shaft
Electrical torque
Exciter control voltage (generator terminal voltage)
Sensed terminal voltage
Voltage regulator reference voltage (determined to satisfy initial conditions)
Excitation system stabilizer output
Voltage regulator output
Maximum and minimum voltage regulator outputs
Exciter block constant
xiii
Exciter block time constant

Exciter saturation function
Excitation control system stabilizer gain
Excitation control system stabilizer time constant
Exciter output voltage and generator field voltage
Lead-Lag block output
Voltage regulator time constants
Voltage regulator gain
Speed Reference (determined to satisfy initial conditions)
Speed droop
Temporary droop
Governors time constant
Filter time constant
Servo time constant
Hydraulic systems time constant
Gate opening
Water flow
Water head
No load flow
Turbine gain
Turbine damping
Gate velocity limits
Fuel Flow to the combustion chamber
Fuel valve opening
Maximum and minimum valve position outputs
Turbines Mechanical Power
Exhaust temperature load
Ambient temperature load limit
Governor time constant (Gas Turbine)
Combustion chamber time constant (Gas Turbine)
Exhaust gas measuring system time constant (Gas Turbine)
Temperature control loop gain
Governor time constant (TGOV1)
Boiler and re-heater dynamics time constant
Valve position output
or
Simulation time step
xiv
List of Terms
Specied Active and Reactive Power Bus Load bus
PQ
Specied Active Power and Voltage Bus Generation bus
PV
PSS/E
TM
Power Systems Simulator for Engineers
IEEE
Institute of Electrical and Electronics Engineers
ATP/EMTP
Alternative Transients Program / ElectroMagnetic Transients Program
MATLAB
Matrix Laboratory
PSS
Power System Stabilizer
xv
List of Programs
MATLAB
PSS/E
TM
Microsoft Word
Microsoft
PowerPoint
MathWorks high-level language environment

Power Systems Simulator for Engineers
Microsofts word processor
Microsofts slide presentation and draw tool
xvi
1. Introduction
Contents
1.
Introduction ...................................................................................................................... 1
1.1.
Context and Motivation ............................................................................................. 2
1.2.
State of the Art .......................................................................................................... 3
1.3.
Objectives.................................................................................................................. 5
1.4.
Dissertation Outline ................................................................................................... 6

1
1.1. Context and Motivation

The scope of this dissertation is the dynamic and transient analysis of the response to severe
disturbances in power systems. These types of disturbances cause changes in the state of
practically all the generation equipment as well as in the network itself, hence the importance in
this kind of analysis.
Much has happened since the first complete electric system built by Thomas Edison in 1882,
which consisted of a DC system used to power incandescent lamps. One of the most important
aspects on the history of electric power systems was the overlap of AC to DC systems. In these
early years active discussions over which should be the standard system type were formed. The
key actors in this discussion were Edison, who argued in favour of DC systems, and George
Westinghouse, who along with Nikola Tesla preferred the AC systems. It was precisely after the
emergence of polyphase systems developed by Tesla, and the conception of transformers
through L. Goulard and J.D. Gibbs that AC system became so attractive. With the turn of the
century AC systems became once and for all dominant, shaping the systems that are still used
today [6].
During this time, power systems have been developed, increasing the transmitted powers, as
well as the distances between generation and consumption. The standardization of frequency
led to the possibility of interconnections, which in their turn provided an enhancement in system
security, since the junction of the networks results on mutual emergency assistance, and an
improvement in economy of operation [6].
These systems nowadays have as their main objective the transportation of electrical energy,
from the generation centre to the consumption points, in a continuous and reliable manner.
Thus, these systems must be capable of handling small and large disturbances with which they
are confronted, and then be able to operate satisfactorily. For example, as power systems
depend heavily on synchronous machines for the generation of electrical power, a satisfactory
system operation is obtained when all synchronous machines are in synchronism. To make this
possible, a generation-consumption balance should be maintained at all time.
Power system analysis techniques have been clearly modified with the development of digital
computation. Combining the theoretical and empirical knowledge obtained over the years with
the new computing capabilities, it became possible to simulate and analyze systems and their
response to occurred disturbances, in a more rigorous and precise way. Thenceforward many
commercial simulation software packages emerged, and have been used by engineers for
analyzing and designing power systems. However, due to the commercial nature of these
programs, the access to the dynamic models, as to its components, are hindered, making it
impossible for them to be consulted or personalized by the user. From an academic point of
view, this restrains the learning processes since the construction of the dynamic models and the
program procedures are important features when the intrinsic study of dynamic models and
simulations is required.
This dissertation aims to further expand a dynamic simulation program which has been under
development in previous works performed by former I.S.T. master students. This simulation
package is intended for educational use, while accomplishing an approximate or even similar
level of precision when compared to existing commercial packages. This work has a particular
interest for the simulation of existing AC systems of realistic size, which requires a special focus
on the generator control systems speed governing and excitation system as they are crucial
for the stable operation of large networks.
1.2. State of the Art

The stability of a power system can be seen as a single problem; however, its study cannot be
accomplished without further considerations, since the instability of a system can take many
forms and be a consequence of numerous factors. Therefore, it is necessary to divide the study
of dynamic stability of the power system into several categories, each one with its own
characteristics. A classification proposed by [6] is based in the following considerations:
The physical nature of the resulting instability;
The size of the considered disturbance;
The devices, processes, and time span that must be taken into consideration in order
to determine stability;
The most appropriate method of calculation and prediction of stability.
The consideration associated with the size of the disturbance, leads to two types of stability
analysis
Small-signal stability, which is due to small disturbances - These refer to the

disturbances caused by, for example, small variations in load or generation that cause a
response by the governor; and
Transient stability, which is a consequence of severe disturbances Short circuits,

voltage collapses, etc. The main concern of these studies is the loss of synchronism in
a power system.
The principal method for analyzing the dynamic stability is digital simulation, with the aid of
computers, due to its ability and flexibility. However, because of the large size of the existing
interconnected networks, even with the evolution of computational resources, it is necessary to
conduct such studies with special care.
One way to achieve economy of resources is to differentiate the nature of the simulated
phenomena, within the two configurations listed above. An arrangement of dynamical
phenomena based on its time scale can be seen in Figure 1.1.
Figure 1.1 Time ranges of dynamic phenomena (Source: [8])
This classification shows that different levels of detail are required, due to the time scale of each
problem. The time constants on the lower side of the time-scale are very small and therefore,
smaller time-steps are needed in order to prevent numerical errors. As we move to the upper
side of the time-scale, this kind of precision requirement decreases, allowing the use of bigger
time steps. Observing Figure 1.1, it is also noticed that a separation between network elements
and generating unit components can be made. For example, the study of electromagnetic
transient phenomena requires great detail in the transmission line equations, whereas for the
models of the turbine there is no need for a very high detail. Considering a small load variation,
the model of the generator and governor need a level of detail far higher than the models of the
wave equation. Depending on the nature of the considered disturbance, different simulation
software packages may be used.
For higher frequencies phenomena, programs with a detailed model of the network elements
and with the capability to perform simulations with small time-steps are recommended. An
example is ATP-EMTP software package, which is specialized in electromagnetic transient
models.
In the simulation of large electrical power systems in the range of transient and long term
stability, the interest frequencies are smaller; hence the extremely small time steps are no
longer required. EUROSTAG and PSS/ETM are two of the mainstream software packages used
for the analysis of this kind of phenomena
The MATLAB program under development, aims to simulate severe disturbances, specifically
three phase short circuits and the subsequent removal of the line to isolate the fault. This kind of
dynamic phenomena is in the band of transient stability with low frequencies and therefore
PSS/E
TM
simulation software was chosen to validate the implemented dynamic models on the
MATLAB program.
1.3. Objectives
The main ambition of this dissertation is to implement a student-grade software package
capable of simulating electric power systems with realistic size, for transient stability studies.
Thus, the final aim is the simulation and analysis of a 57-Bus IEEE network.
The work consists on the expansion of a MATLAB environment simulation package, which has
been developed in past works by former IST students, giving the program the ability to deal with
large networks, which was not possible until the completion of this dissertation1.
As the size of the networks grows, so does the difficulty in controlling the power systems.
Hence, we will review the existing generator models and mainly the control systems (voltage
and governor regulators), in order to further complete these models, so that an accurate
treatment of the systems is possible. It should be noted that, due to the large size of the
networks, these control models are essential for the stability of the system after a fault
occurrence.
New dynamic models, representing common components found in most systems will also be
added to the MATLAB software package. This way, the construction of the models in differential
form and the subsequent conversion to an algebraic shape, necessary for digital computation, is
exemplified.
The simulation tests are carried out on the last stage of the dissertation. Firstly, the newly
implemented models, as well as the upgraded ones, are validated using a simple network that
consists of two buses, a generator and a load. In order to confirm the accuracy and precision of
these models, the simulation results are presented side-by-side with the obtained results from
PSS/ETM, using the same conditions and input files in both programs. Afterwards, the target
simulation the 57-Bus network is presented. All the library models are used on this
simulation case, providing an analysis, not only of the transient stability of the power system,
but also of the different dynamic responses presented by each model.
1
The larger simulated network in the previous version of the software was a nine bus system.
Table 1.1 and Table 1.2 show the studied dynamic models in this dissertation. On the left side
are the models names and on the right side is their respective PSS/E
TM
nomenclature. These
designations are used for the models implemented in the MATLAB software as well.
Table 1.1 Revisited dynamic models
Dynamic Model
Nomenclature
Round rotor synchronous machine with exponential saturation in d-q axis
GENROE
Round rotor synchronous machine with quadratic saturation in d-q axis
GENROU
Salient-pole synchronous machine with exponential saturation in d-q axis
GENSAE
Salient-pole synchronous machine with quadratic saturation in d axis
GENSAL
IEEE Type 1 Exciter, Excitation control system
IEEET1
Steam Turbine and Governor with reheat
TGOV1
Hydraulic Turbine and Governor
HYGOV
Table 1.2 Newly implemented models
Dynamic Model
Nomenclature
Type DC1A Exciter, Excitation control system
IEEEX1
Gas Turbine and Governor
GAST
1.4. Dissertation Outline

This document is organized in six main chapters. In order to better understand the structure of
this dissertation, an overview of each one of these chapters is made in the following
descriptions.
Chapter 1 - Introduction contextualizes the work. The interest in dynamic simulations, as well
as the motivations and aimed objectives of the dissertation are given. A concise approach of the
state of the art is also done.
Chapter 2 - Power System Simulation Software gives an insight of the simulation software.
Firstly, the procedures and the processes of the program are exposed and afterwards a few of
the most important changes made to the software are presented.
Chapter 3 - Dynamic Models Description describes the improved and the newly added
dynamic models, while giving an overview of the importance of the control systems in regulating
the parameters of the network.
Chapter 4 - Differential-Algebraic Model starts by presenting a newly added integration
algorithm - the Modified Euler-Cauchy. Afterwards, this chapter presents the dynamic models in
differential and algebraic form, required to compute the state equations. Finally, the initial
conditions for the dynamic simulations are also computed.
Chapter 5 - Simulation Results presents the results and the analysis of the simulations
performed to validate the models and of the main simulation of this dissertation.
Chapter 6 - Conclusions summarizes the work and the achieved conclusions. This chapter
also offers some suggestions for the further expansion of the simulation software.
2. Power System
Simulation Software
Contents
2.
Power System Simulation Software ................................................................................... 9

2.1.
Simulation Software Algorithm ................................................................................ 10
2.1.1.
Simulation Data Acquisition ............................................................................. 10
2.1.2.
Power Flow Computation................................................................................. 10
2.1.3.
Modelling of the Network Equations ................................................................ 12
2.1.4.
Computation of the Initial Conditions for Dynamic Simulations ........................ 12
2.1.5.
Computation of the Matrices of the Algebraic Dynamic State Equations .......... 13
2.1.6.
Dynamic Simulation ......................................................................................... 13
2.2.
Simulation Software Improvements and Modifications ............................................ 15
2.2.1.
Reading of the Dynamic Data Files ................................................................... 15
2.2.2.
Presentation of the Results .............................................................................. 17
2.2.3.
Generator Reactive Power Limits ..................................................................... 18

9
This Chapter is reserved for the presentation and explanation of the developed simulation
software. The whole simulation procedures and the aspects related with the program are
reviewed, in order to fully understand how the software package works. Some of the
improvements and modifications made to the program are then described. These upgrades
were necessary for the accomplishment of one of the plotted objectives for this dissertation
the simulation of realistic systems with realistic dimensions.
2.1. Simulation Software Algorithm

This Section shows the steps and the routines of the program. For an easier comprehension, a
basic scheme of the simulation process is displayed in Figure 2.1.
2.1.1. Simulation Data Acquisition

The first step consists of reading the input files. There are two types of files; the power flow data
files and the dynamic data files. The first ones contain information regarding the network, that is,
bus data, branch data, generation requirements and power demand data. Most of this
information is used to build the admittance matrices and also for the computation of the power
flow. The dynamic data files hold the information regarding the parameters of the synchronous
generator and control systems.
The power flow files have the extension *.raw, while the dynamic files are terminated with *.dyr.
These extensions were selected so that the MATLAB program files are compliant with PSS/ETM
files. Evidently the extensions are not the only similarities between the two program files, as
their structure is also intended to be the same.
2.1.2. Power Flow Computation

While the main objective of the simulation program is to obtain the transient solution of the
power system after a fault, the calculation of the load flow is an essential part of the process.
From the power flow computation we obtain the results that are used for the generators and
loads conversion, as well as for the calculation of the dynamic initial conditions, which is crucial
for the accuracy of the final outcome. In order to compute the load flow, the Newton-Raphson
numerical solution is used. This is the reference in load flow solution methods which,
independently of the network size, converges in three to five iterations as long as there is no
reactive limits violation, which would oblige a bus type adjustment [9].
10
Figure 2.1 Simulation process flowchart
11
2.1.3. Modelling of the Network Equations

The next step is the modelling of the network equations in the form of an admittance matrix,
Although this matrix was already built in the load flow calculation, a new matrix must be
determined for the transient solution, because the generators and the load demands must be
included in this calculation.
Regarding the load conversion, the constant admittance method was adopted. This method
considers that the loads can be converted into pure equivalent impedances, by using ( 2.1 ),
and that they are later added in the network admittance matrix
, as if it was a real admittance.
( 2.1 )
For dynamic simulations, the generators are also converted. Each generator may be modelled
as an equivalent impedance, the subtransient impedance, which is the impedance used to
determine the current during the first cycle after the occurrence of a fault. As this parameter is
accounted for in both *.raw and *.dyr files, caution must be taken to insure that it is equal in both
files, to prevent numerical errors.
After the conversion of the generators and loads, the admittance matrix is recalculated. Then, in
order to decrease the computational effort, the network is reduced by using the Internal Node
method. This technique reduces the admittance matrix by eliminating all the network nodes with
the exception of the internal generator nodes. It is not given a full explanation of the method,
since it is out of this dissertations scope. However, more detailed information about this topic
can be found in [9] and [10].
2.1.4. Computation of the Initial Conditions for Dynamic Simulations

One of the most important steps in the simulation procedure is the computation of the initial
values of the dynamic state variables. This, together with the load flow results, acts as a
checkpoint before entering the dynamic simulation stage. If the initial values of the state
TM
variables differ from the ones obtained by PSS/E , we have an indication that something is
wrong, and that further improvement in previous stages is required. The computation of the
initial condition is specific to the type of machines and control systems being used. Hence only
a straightforward explanation of the proceedings is given here. In a more practical way, Section
4.4 provides the expressions of the initial conditions for each specific dynamic model.
The program follows a method proposed in [8], which suggests the following sequence of steps
to determine the initial values of the state variables.
Step 1 Calculate the generator current in the network frame,
12
Step 2 Compute the initial value of the machines rotor angle,
Step 3 Compute the generator current and terminal voltage in the machine reference
frame,
and
Step 4 Calculate the initial value of the transient emf of the direct axis,
Step 5 Compute the initial value of the transient emf of the quadrature axis,
Step 6 Compute the initial value of the Electrical Field Voltage,
Step 7 Compute the initial values of the mechanical torque and speed,
and
These steps are related with the state variables and fixed inputs of the synchronous generator.
If the excitation and governor systems are coupled with the generator, the initial values of their
dynamic-state variables, as well as their fixed inputs computation is also required. Chapter 4
details dynamic differential-algebraic equations of the models, thus giving, as mentioned before,
the complete expressions of the initial conditions for each one of the dynamic models.
2.1.5. Computation of the Matrices of the Algebraic Dynamic State

Equations
One last step remains before entering the dynamic simulation the construction of the matrices
of the algebraic state equations, which are used in the digital numerical integration. These
algebraic equations are derived from the differential expressions that represent the dynamics of
each one of the systems included in the generator group. Section 4.3 - Models in Algebraic
Form, fully presents the arrangement of the algebraic state equations of each dynamic model.
2.1.6. Dynamic Simulation

Finally, after the preliminary calculations the dynamic simulation is at hand. As it is known, this
type of reckoning uses a discrete method, due to the inherent digital nature of computers. The
simulation process is conducted in various integration steps and, in each interval the solution of
the variables is computed.
The simulation starts at the time
; the systems states are according to the initial conditions
determined previously. In each time-step the first task is the computation of the algebraic
equations. These correspond to the representative equations of all the system components
(generator, regulating systems and network) that are not differential and therefore, are apart
from the numerical integration. The synchronous machine algebraic variables are composed by
the subtransient internal voltage,
, and the injected currents on generator buses,
13
depends on which synchronous generator is used (round rotor synchronous generator or salient
pole synchronous generator).
is established by using the reduced admittance matrix and
( 2.2 )
After obtaining the injected currents, the network voltages (the voltages in every bus) are
computed using eq. ( 2.3 )
( 2.3 )
With the results of equations ( 2.2 ) and ( 2.3 ) the complex power is obtained by using ( 2.4 ),
and consequently the generated active and reactive powers are computed.
( 2.4 )
( 2.5 )
( 2.6 )
The next step is the computation of the algebraic variables associated with the dynamic control
models. As this Chapter is intended to generally overview the program, these equations are not
displayed here. Instead, they are shown and well identified in Section 4.2, where these models
are presented.
After the computation of the algebraic equations, the dynamic simulation checks if there is a
network topology change (which corresponds to a fault). If so,
is changed and consequently
reduced. If the network topology is unaffected we jump to the computation of the machine state
equations parameters that need to be computed in every time step2. Everything is now set to
establish the computation of the state variables using the numerical integration algorithm.
In this simulation package, there are two options regarding the used integration method; one
that was previously implemented the Fourth Order Runge-Kutta integration algorithm and
another, implemented during the developed work in this dissertation the Modified EulerCauchy integration algorithm which is a more proximate method with the one used in
TM
PSS/E . The description of this method is given in Section 4.1.

Time is then incremented by one time-step, and a comparison between the present and the
maximum specified times is made. If the maximum time is not reached, the simulation
continues. Otherwise the simulation is concluded, and the results are plotted.
As we will see in Section 4.3, the algebraic state equation is divided into constant and non-constant
parameters
14
2.2. Simulation Software Improvements and Modifications

This dissertation aims to look into the most basic details of the MATLAB program, in order to
improve it in all aspects performance, robustness, technical issues, compatibility. The
previous version of the software was not designed to deal with large networks, and therefore
some adjustments and upgrades had to be made for the completion of the outlined goals. Some
of these modified features, related with the structure of the software, are presented in this
Section.
2.2.1. Reading of the Dynamic Data Files

The MATLAB simulation software is intended to be compatible with the mainstream software
TM
package PSS/E . Consequently, the used input data files should be exactly equal to the ones
used by PSS/ETM. Thus, the function that read the dynamic data files was changed, in order to
consider spaces for the separation of data, instead of commas3.
One of the possible displays of the PSS/ETM dynamic data files is displayed next:
1 'GENROE'
3.00000
0.60000
0.60000E-01
1.00000
0.50000E-01
1.00000
1.40000
1.35000
0.30000
0.30000
0.10000
0.00000
0.00000
0.02300
200.000
0.84000
99.9999
-99.9999
1.00000
0.30000
0.67000E-01
1.00000
0.00000
0.00000
0.00000
0.00000
0.00000
0.50000E-01
0.40000
9.50000
0.00000
1.00000
0.00000
1 'IEEET1'
6.50000
'TGOV1'
0.40000
By observing the previous data sample of a *.dyr file, we see that the data is organized in fixed
columns. The new function starts by predetermining the length of eight arrays by allocating the
number of characters from the beginning of the string (or row) to the last character of each one
of the seven columns presented in the file. Then, the size of
difference between the last element of
can be found through the
and the last element of
As said before, the size of these columns are predetermined, which is in contrast to what
happens in the original PSS/E
TM
dynamic files. This method, although somewhat restrictive
since it imposes the data to have a predefined structure, leads to a simplification from a
programming point of view. This does not compromise the initial objective, since the data files
TM
are in fact compatible with the dynamic data files used by PSS/E .
In the former version of the MATLAB program, dynamic files were read as the *.raw file, in which the data
separator is a comma
15
Table 2.1 shows the chosen array distribution.

Table 2.1 Array distribution
length 0
Array length
(character count)
0
length 1
05
length 2
0 15
length 3
0 19
length 4
0 32
length 5
0 45
length 6
0 58
length 7
0 70
length interm
0 44
Array name
And therefore the columns distribution is as shown in Table 2.2.

Table 2.2 Column distribution
column 1
(1 5)
column 2
(6 15)
column 3
(16 19)
column 4
(20 32)
column 5
(33 45)
column 6
(46 58)
column 7
(59 70)
After the column allocation, each row of the *.dyr file is read once at a time, constituting an array
of characters - or string. By using the previously defined columns, we can separate each data
element given by a string and then address it to the respective data field variable.
In practical terms, this means that the first column in the first row (BUSID bus identification)
has 5 assigned characters, the second column (Model Name or machine name) has 10
allocated character, the third element (ID Machine ID) has 4 assigned characters, the fourth
element has 13 assigned elements, and so forth. It should be noticed that on the second and
third rows of each dynamic model, the first element is a combination of column 1, column 2 and
column 3. Therefore, the first column of these rows has 19 allocated characters.
This function greatly facilitates the comparison between the results obtained by MATLAB and
PSS/ETM because the dynamic files no longer need to be converted, as they were before. This
16
conversion was handmade and, in case of large networks, it was an unnecessary and a timeexpendable task.
2.2.2. Presentation of the Results

The plotting of the results is an extremely important feature, as it shows graphically the
response of the system to a fault occurrence.
One of the main limitations on the preceding version of the program was that it was designed to
accommodate only three specific cases a 2-Bus network case, a 3-Bus network case and a 9Bus network case. To plot the obtained results, each simulated case needed a specific function.
In addition, in each function, the variables were plotted one by one in a non cyclic manner (a
line code for each plotted variable). Therefore, if any other given case was to be simulated, the
program could not successfully complete the simulation, because the presentation of the results
was case dependent. This plotting method was time consuming and, besides, it added
unnecessary code for big networks; for example, the 57-Bus network has more than one
hundred variables to be plotted, which implied more than one hundred line codes, one for each
variable. As the software is intended to deal with any given network, independently of its size,
the plotting method was changed in order to automate it. So, a new function, which no longer
requires additional code each time a new case is tested, was created.
One other feature was added to the presentation of the dynamic results. As new and larger
networks were tested, it was found that the plotting of every variable was consistent with
PSS/ETM, with the exception of the rotor angle, unless a reference was given to this variable.
Large power networks have a multimachine configuration. The multimachine models used in
this software have at least one more differential equation than is needed to solve an m machine,
n bus problem, because every rotational system must have a reference for angles [8]. There are
various types of angle references. The default arrangement for the angle reference is the
absolute angle, which is not exactly a reference, since it gives the absolute value of each
machines angle. The results obtained by using this configuration, did not complied with the
results given by PSS/ETM. To solve this problem it was decided to force a reference in the angle
representation. The chosen method was to define all the angles relative to the angle of the
balance bus machine. Using this method of representation, the reference angle is zero
throughout the entire simulation while the rotor angles of the other generators are given by:
After applying this method to represent the rotor angles, all the results were similar to the ones
retrieved from PSS/ETM, confirming the consistency of the results.
As said, the number of differential equations needed to solve the dynamic solution could be
reduced. However, this was not done since it would imply a complete change in the numerical
17
calculations structure, which was not the objective. So, it was decided to implement this
reference only in the plotting of the results.
2.2.3. Generator Reactive Power Limits

Another implemented feature was the control of the generated reactive power in the load flow
computation. As stated in [3], generator reactive power limit is a key factor in voltage instability.
The power system must operate within acceptable voltage limits and reactive power flows can
be responsible for substantial voltage changes across the system. Therefore, it is important that
reactive power is maintained at a balanced level.
The function that emulates this control is inserted in the Newton-Raphson algorithm. After
computing the injected powers, angles and voltages in all buses, the program verifies the
reactive powers are within limits. Figure 2.2 shows a basic sketch of the implemented function.
Figure 2.2 Reactive Power limit verification function
The implemented algorithm starts by checking if the generated reactive power on a generation
bus (or PV bus) is within the generator specified limits. If not, the bus type changes from PV to a
false PQ (or false load bus) and the generated reactive power is specified with the limit value,
or
.
18
In the next iteration the generator bus is seen as a false PQ bus. At this time, the function
computes the injected reactive power in this bus by using its specified voltage, and, with this
result, the generated reactive power is calculated and compared with the reactive limits of the
generator. If
is within limits, the bus is once again classified as a PV bus. Otherwise, it
continues to be a false PQ bus, and the generated reactive power continues to be specified by
the generator breached limit.
This Section presented a few of the most relevant modifications made in the simulation
software, although many others have been made. Here, only the features closely related with
the structure of the program and the preliminary calculations were given. The following Sections
present the specific dynamic models and dynamic simulation upgrades. This separation was
made, since the main scope of this dissertation is in fact the dynamic simulation and therefore,
this topic deserves special attention.
19
20
3. Dynamic Models
Description
Contents
3.
Dynamic Models Description ........................................................................................... 21

3.1.
Round Rotor Synchronous Generator ...................................................................... 22
3.2.
Excitation Control Systems....................................................................................... 23
3.2.1.
Excitation Control System IEEE Type I, IEEET1................................................... 25
3.2.2.
Excitation Control System Type DC1A, IEEEX1 .................................................. 26
3.3.
Governor Control Systems ....................................................................................... 27
3.3.1.
Hydro-Turbine Governor, HYGOV ..................................................................... 27
3.3.2.
Gas-Turbine Governor, GAST............................................................................ 29

21
This Chapter describes some of the dynamic models used in this dissertation. At first, an
introductory presentation of a synchronous generator dynamic model is given. The four existing
4
generators in the library of the software package were already implemented in the previous
versions. However, one of these models is exemplified in order to give an insight about the
dynamic arrangement of the generators. This insight is considered important, since these
models are attached with the control systems shown in the following sections and therefore, is
crucial for the understanding of the dissertation. The round rotor synchronous generator is the
chosen generator to be described, because a bug was encountered on the initial conditions
computation of GENROU. This bug was due to incorrect code implementation and it prevented
the correct calculation of the initial values of the state variables. In some cases it even caused
an endless loop. Afterwards, the control systems (excitation and governor systems) and their
dynamic models are described. The portrayed control models are the excitation systems
IEEET1 and IEEEX1, and the turbine governor systems HYGOV and GAST. The steam-turbine
governor TGOV1 was also reviewed in the developed work on this dissertation. However, due
to the lack of space, it is not described. Details about this model can be seen in [10] and in [17].
3.1. Round Rotor Synchronous Generator

The electromagnetic model of the round rotor synchronous generator is displayed in Figure 3.1.
This representation can be seen in [10] and is the one used in PSS/ETM as well. It should be
noted that the electromagnetic model represented in Figure 3.1 is the same for both GENROU
and GENROE, which only differ on the magnetic saturation function. GENROU model
represents a round rotor synchronous generator with quadratic saturation in both axis, while
GENROE model represents a round rotor synchronous generator with exponential saturation in
both axis.
The physical aspects of these models are not described on this dissertation. Excluding the
computation of the initial conditions for GENROU, no significant changes were applied to the
generator models, as their implementation in the previous work was accurately completed. A
more comprehensive portrayal of these synchronous generators (as well as of the Salient Pole
Synchronous Generators) can be found in [10], [8] and [6].
Round Rotor Synchronous Generator (with quadratic and exponential saturation in both axis) and
Salient-pole Synchronous Generator (with quadratic saturation quadratic saturation on d-axis and
exponential saturation in both axis)
22
Figure 3.1 - Electromagnetic model of the round rotor generator, GENROU and GENROE (Source: [15])
3.2. Excitation Control Systems

Excitation control systems are responsible for the voltage regulation of the power network.
These systems should maintain the machine terminal voltages between specified and workable
limits. Outside these limits, particularly for long periods, these voltages adversely affect the
performance of the generator, possibly harming it. The excitation systems accomplish this
regulation by controlling the generator input voltage,
. These systems also assure the
stability of the voltage.

The various excitation subsystems of synchronous generators are indicated in Figure 3.2.
23
Figure 3.2 Synchronous machines excitation subsystems (Source: [6])
For the simulated cases in this dissertation, it is considered that the terminal voltages are not
remotely controlled; therefore the load compensator is not implemented. Also, as no Power
System Stabilizer models were implemented, the upcoming excitation models do not consider
PSS feedback. The limiters and protective circuits include various types of control and
protective functions which ensure that the physical limits of the exciter and of the synchronous
generator are not exceeded. The limits considered in the implemented excitation models are the
voltage regulator limits.
There are three distinctive types of excitation systems:
Type DC Excitation Systems
Type AC Excitation Systems
Type ST Excitation Systems
This dissertation concentrates on the DC excitation Systems.
Type DC Excitation Systems are the older ones. These utilize a shaft-driven dc generator as the
source of the excitation system. With the advent of power electronics they were surpassed by
Type AC and ST systems, and therefore few new synchronous machines are being equipped
with Type DC exciters. Nevertheless many of these systems are still in service [5]. The following
subsections describe two variants of this type of excitation control implemented in the MATLAB
software the excitation control systems IEEE Type I (IEEET1) and Type DC1A (IEEEX1).
24
3.2.1. Excitation Control System IEEE Type I, IEEET1

A simplified version of this system was already implemented in the previous version of the
program and been described in [10]. However, after revisiting this model, it was decided to
include some missing components (as the terminal voltage measurement, and the voltage
regulator limits), and to improve the general performance of the model. Figure 3.3 shows the
respective block diagram of IEEET1.
Figure 3.3 Block diagram of the IEEET1 dynamic model (Source: [13])
The terminal voltage of the generator,
, is the input of a voltage measurement system (or
terminal voltage transducer) that acts as a filter with time constant

some systems
is very small and may be considered zero.
The first summing point compares the reference voltage,

transducer,
. It should be noted that for
, determining the voltage error,
, with the output of the voltage
. The value of
is constant throughout the
simulation and is determined on the initial conditions computation, considering that
is with its
proper value at t=0 [2].

The second summing point combines the voltage error input with the excitation stabilizing
feedback voltage and produces the input signal for the regulator. The excitation stabilizing loop
is represented by a differential feedback block which corresponds to a powerful tool to stabilize
a closed loop system [9]. This loop is used to eliminate the static error and to assure a fast
dynamic response of the system.
The voltage regulator amplifier has a gain
and a time constant
. These parameters are
shown incorporating a non-windup limit configuration, typical of saturation or amplifier power

supply limitations [2]. A discussion of windup and non-windup limits is presented in Appendix A.
These limits are imposed so that the output of the regulator cannot exceed practical limits.
25
The output of the regulator, , is used to control the exciter. This system can be either a
separately excited or a self-excited shunt field and the value of
reflects the setting of the
shunt field rheostat. In the case of a separately excited exciter a value of

The input of the exciter block is the combination of
is used [12].
and a term given by
represents the saturation factor of the exciter (Appendix B), and is a nonlinear function
of the exciter output voltage,
It should be noticed that excitation control systems are fast and have small time constants,
sometimes even considered zero, due to its intrinsic electronics. This fact must be taken into
consideration when designing and implementing these models. The selection of the simulation
time-step must be made with care in order to correctly deal with these systems. Regarding the
time constants of the model, if the time-step is too big, numerical errors may appear. The
selected time-step should always be at least 4 times smaller than the smallest time constant of
the dynamic model, in order to prevent numerical errors in the integration of the state variables.
3.2.2. Excitation Control System Type DC1A, IEEEX1

Figure 3.4 shows the block diagram of the excitation system type IEEEX1.
Figure 3.4 Block diagram of the IEEEX1 dynamic model (Source: [13])
After a close observation and manipulation of the block diagrams displayed in Figure 3.3 and
Figure 3.4, we see the resemblance between IEEET1 and IEEEX1. In fact, the only difference
between the two diagrams is the introduction of a lead-lag block in the voltage regulator of
IEEEX1. This block uses time constants
and
to model equivalent time constants inherent
to the voltage regulators, that were not accounted for in the IEEET1 model. Due to the similarity
between the two models, the description presented in Subsection 3.2.1 is also applied to
IEEEX1 system. The modelling of the differential-algebraic equations shown in Chapter 4 sets
the numerical differences between the two excitation systems.
26
3.3. Governor Control Systems

The frequency of a power system is a global quantity and should remain nearly constant,
typically
of nominal frequency [9], for the stable operation of the network. The frequency
control of the system ensures the constancy of speed of the synchronous (and induction)
motors, which is particularly important for the satisfactory performance of the generating units.
The frequency regulation is closely related with the balance between production and
consumption of active power. Therefore, a change in power demand at a certain point of the
network is reflected throughout the system by an adjustment of the frequency. Governor control
systems ensure that generators satisfy the changes in demand so that the active power balance
is maintained.
Each generation unit is provided with a speed governor, which assures the called primary
control. A basic scheme of a speed governor is shown in Figure 3.5.
Figure 3.5 Speed Governor Basic scheme (Adapted from [9])
When load demand changes, the generated active power vary, leading to mismatches between
mechanical and active powers, which result in variations on the speed of the machine. The
governor measures the rotating speed of the unit and compares it with the reference. Based on
the resultant error, the admission valves or gates will open or close in order to increase or
decrease the mechanical power so that the mismatch gradually disappears.
3.3.1. Hydro-Turbine Governor, HYGOV

HYGOV is one of the revisited models. The performance of this model is intended to be
improved by implementing missing components such as gate opening limits and gate velocity
limits. This section briefly describes the model while a more detailed explanation of the physical
characteristics can be found in [10] and [16].
The block diagram and the parameters of the dynamic model are respectively presented in
Figure 3.6 and in Table 3.1.
27
Figure 3.6 Block diagram of the HYGOV dynamic model (adapted from [17])
Table 3.1 Parameters of the HYGOV dynamic model
Parameter
Representation
Units
Speed reference
Speed deviation
Permanent droop
Temporary droop
Governors time constant
Filter time constant
Servo time constant
Hydraulic systems time constant
Gate opening position
Desired gate opening position
Water flow
Water head
No load flow
Turbine gain
Turbine damping
The model can be separated in two subsystems, the governor and the hydraulic systems.
The two inputs of the governor system are the speed reference,
machine,
, and speed deviation of the
is obtained in the computation of the initial conditions, and
is acquired by
the speed measurement system of the machine. The governor system finds the speed error
28
through the comparison between

is composed by the addition
and the output of the feedback path. This feedback path

. The second term consists of the combination of the
speed droop, , with the desired gate position, , which creates an equivalent speed regulation
under steady state conditions.
The speed error is then filtered, and later is used as input on the governor block, which provides
the desired gate position. This variable is the input of the servo motor, which determines the
gate opening value.
The governor block has two limits to take into consideration. One portrays the maximum and
minimum limits of the gate opening values with a non-windup configuration, and the other
represents the gate velocity limit. In Subsection 4.2.4 this matter is further discussed.
The inputs of the hydraulic system are the gate position and, in case the speed damping is
considered, the speed of the machine. The output of this system is the mechanical power
delivered to the generator. Since this dissertation only makes a brief review of the HYGOV
model, the parameters and computation methods of the hydraulic system are not presented,
because it is a very specific topic. [16] and [1] provide a detailed description about the hydraulic
system.
3.3.2. Gas-Turbine Governor, GAST

Gas turbines have become increasingly popular mainly because of their lower greenhouse
emission and higher efficiency when compared with other turbines, particularly when connected
in a combined cycle setup [4].
A gas turbine usually consists in a compressor, a combustion chamber and a turbine operating
under the Brayton cycle. It can operate in both open and closed systems, being that the first one
is the most currently used. Figure 3.7 represents this type of operation.
Figure 3.7 Open system gas turbine
29
In an open system configuration, the working fluid (in this case air) is conducted into the
compressor, where it reaches high pressures. Afterwards, the compressed fluid enters the
combustion chamber and, together with the fuel, originates a combustion which produces high
temperature exhaust products. These products expand in the turbine and are later expelled to
the atmosphere. This is the process from which the mechanical energy is produced.
In stability studies, the model of a gas turbine usually models load-frequency control,
temperature control and acceleration control [7]. There are various dynamic models proposed
for stability analysis. This subsection presents the GAST model, shown in Figure 3.8. This is
one of the most commonly used dynamic models due to its simplicity and ability to represent the
principal dynamic characteristics of industrial gas turbines driving generators connected to
electric power systems [16]. Table 3.2 pinpoints the parameters of GAST dynamic model.
Figure 3.8 - Block diagram of the GAST dynamic model (Adapted from [17])
Table 3.2 Parameters of the GAST model
Parameter
Representation
Units
Speed droop
Fuel Flow to the combustion chamber
Fuel valve opening
Maximum valve position
Minimum valve position
Turbines Mechanical Power
Exhaust temperature load
Ambient temperature load limit
Governor time constant
30
Combustion chamber time constant

Exhaust gas measuring system time constant
Temperature control loop gain
In this model, speed variations are expected to be small (approximately 5%).
The block diagram represents a forward path consisting in a valve controller block with a time
constant of
and the combustion chamber block with a lag time
. In the diagram, the load-
frequency and temperature control models are also represented.
is the exhaust gas
measuring system time constant.

The load-frequency control is the main control loop and consequently is active in normal
operation conditions. The input of this control is the speed deviation
. The fixed droop, , is
used to regulate the power generator. As seen in Figure 3.8, in load-frequency regulation the
gain is the inverse of the permanent droop. Using this technique, the reference speed of the
governor is reduced as load increases thus providing a stable operation [4].
The temperature control serves as a protective function. As load demand increases, in normal
operation conditions, load-frequency control forces the gas turbine to raise its output power.
Therefore, the amount of fuel delivered to the combustion chamber is increased, rising the
exhaust products temperature. If this temperature is higher than a reference the design
ambient temperature of the turbine the fuel flow is halted to prevent damage to the turbine.
GAST model uses a simple method to set this behaviour. At first, the load limit path measures a
power proportional to the turbine exhaust temperature and then compares it with the ambient
temperature load limit. If the temperature of the system is higher than the reference, then the
temperature control output is lower than the load-frequency output, and therefore takes control
of the gas turbine. This decision is made by the Low Value Gate, which selects the lower value
between both control models, setting the fuel flow to the combustion chamber.
The load limit depends on the ambient temperature in which the turbine is operating. If the
turbine is operating at its design ambient temperature, parameter should be set to unity. If the
turbine operates at temperatures higher than the design ambient temperature, the load limit
should be set to a lower value.
The constant,
, is used to adjust the gain of the load-limited feedback path.
31
32
4. Differential-Algebraic
Model
Contents
4.
Differential-Algebraic Model ........................................................................................... 33

4.1.
Modified Euler-Cauchy integration algorithm .......................................................... 34
4.2.
Models in Differential Form ..................................................................................... 35
4.2.1.
GENROU and GENROE model ........................................................................... 35
4.2.2.
IEEET1 model ................................................................................................... 36
4.2.3.
IEEEX1 model ................................................................................................... 37
4.2.4.
HYGOV model .................................................................................................. 38
4.2.5.
GAST model ..................................................................................................... 39
4.3.
Models in Algebraic Form ........................................................................................ 40
4.3.1.
GENROU and GENROE models ......................................................................... 41
4.3.2.
IEEET1 model ................................................................................................... 42
4.3.3.
IEEEX1 model ................................................................................................... 43
4.3.4.
HYGOV model .................................................................................................. 44
4.3.5.
GAST model ..................................................................................................... 45
4.4.
Initial Conditions for Dynamic Simulation ................................................................ 46
4.4.1.
GENROU and GENROE Initial Conditions........................................................... 46
4.4.2.
IEEET1 Initial Conditions ................................................................................... 48
4.4.3.
IEEEX1 Initial Conditions................................................................................... 48
4.4.4.
HYGOV Initial Conditions .................................................................................. 49
4.4.5.
GAST Initial Conditions ..................................................................................... 50

33
This Chapter discusses the implemented dynamic numerical solutions. At first, a new integration
algorithm to solve the differential-algebraic equations is presented the Modified Euler-Cauchy
integration method. In the following sections, the differential and the algebraic state equations
that represent each dynamic model are derived. In the last stage of the Chapter, the initial
conditions of the models are computed.
4.1. Modified Euler-Cauchy integration algorithm

In the previous version of the developed software, the implemented numerical integration
method was the Fourth-Order Runge-Kutta. Even though being a robust and a stable algorithm,
due to its various intermediate time points to calculate the state at the time
, this method is
TM
not compliant with the one used by PSS/E . In some case, mainly the simulation of large
networks, the results produced by the Fourth-Order Runge-Kutta did not have the same
dynamic behaviour of the results given by PSS/ETM. Hence, a new integration algorithm more
approximate with the one used by PSS/ETM [14] is included, offering a choice for both methods.
The modified Euler-Cauchy is an explicit algorithm, which belongs to the family of the SecondOrder Runge-Kutta method [18], and is given by
( 4.1 )
where
is the state variable,
is the state function and
is the time step.
In opposition to the Fourth-Order Runge-Kutta, which requires four intermediate steps, the
modified Euler-Cauchy is composed of two steps.
Step 1:
( 4.2 )
Step 2:
( 4.3 )
Step 1 moves the state variable a half-step forward to time (
) using the forward Euler
method. Step 2 applies once again the forward Euler method, but at this time using the
intermediate value found in ( 4.2 ).
This way, the modified Euler-Cauchy uses a midway value between
and
. Hence, this
method is an explicit integration algorithm that attempts to share some of the advantages of
implicit methods, by taking midway steps.
34
4.2. Models in Differential Form

This Section derives the differential equations that represent the dynamics of each one of the
models described in Chapter 3. To do the conversion from frequency to time domain, we apply
the inverse Laplace transform to the state variables represented in the block diagrams. The
algebraic equations are also given.
4.2.1. GENROU and GENROE model

Applying the inverse Laplace transform to the state variable presented in the block diagram
illustrated in Figure 3.1, the differential state equations of the round rotor synchronous generator
are obtained.
( 4.4 )
( 4.5 )
( 4.6 )
( 4.7 )
( 4.8 )
( 4.9 )
It should be observed that these equations stand for both GENROU and GENROE. The only
difference lies on the d and q axis saturation factors,
and
, respectively. Due to the
distinct type of saturation of both generators, these saturation factors have different values and
as a consequence, so do the dynamic behaviours of the generators. The saturation factors are
given by:
( 4.10 )
( 4.11 )
35
where
( 4.12 )
( 4.13 )
( 4.14 )
and,
is the saturation function which may represent either a quadratic or an exponential
saturation curve. An insightful explanation of the saturation of the synchronous generator can be found in
[10].
The algebraic equations of the generator models are:
( 4.15 )
( 4.16 )
where m is the number of generators.
4.2.2. IEEET1 model

The differential equations of the IEEET1 model are obtained by applying the inverse Laplace
transform to the resultant equations of the block diagram portrayed in Figure 3.3.
Special attention should be given to the feedback voltage
. In order to simplify the
computation of the differential equations, another model proposed by [8] may be used. This
model defines a new state variable
(called rate-feedback), which has the following form:

( 4.17 )
( 4.18 )
As a result, instead of using
, from now on the used state variable becomes
Taking this into consideration, the representative differential state equations for the IEEET1
model are:
36
( 4.19 )
( 4.20 )
( 4.21 )
( 4.22 )
with the limit constraint of the voltage regulator output
( 4.23 )
As stated in Section 3.2 the regulator limits are of the non-windup type.
Each time
or
exceeds the limit restriction, it is instantaneously fixed with the limit values,
. This implies an iterative computation of the other state variables, as well as of the
algebraic variables, so that they take into account the voltage regulation limitation. This iterative
step requires the integration of the state variables once again. In this process,
is no longer a
state variable becoming a fixed input.
4.2.3. IEEEX1 model

The differential state equations for IEEEX1 are retrieved from the block diagram characterized
in Figure 3.4, and are represented by equations ( 4.24 )( 4.28 ). The considerations regarding
the feedback voltage and the regulator limits, made for IEEET1, should be repeated for this
model. The additional lead-lag block, accounted for in IEEEX1, is the major difference between
the two models. In order to represent the dynamic effects of this block, a new state variable,
, is required.
( 4.24 )
( 4.25 )
( 4.26 )
( 4.27 )
37
( 4.28 )
with the limit constraint of the voltage regulator output

( 4.29 )
4.2.4. HYGOV model

Applying the inverse Laplace transform to the variables shown in the block diagram of Figure
3.6, result the differential state equations ( 4.30 ), ( 4.31 ), ( 4.32 ) and ( 4.33 ).
( 4.30 )
( 4.31 )
( 4.32 )
( 4.33 )
From Figure 3.6, it is also possible to retrieve the algebraic equations of the hydraulic turbine,
given by:
( 4.34 )
( 4.35 )
As already stated in Subsection 3.3.1, there are two types of limit constraints in HYGOV the
maximum and minimum limits of the gate opening values and the gate velocity limit.
The gate position limit imposes that the gate cannot open more than
close more than
, and that cannot
, i.e.
( 4.36 )
At each time step, the desired position is checked. If this variable is not within the limit range, its
value must be fixed, and afterwards, the remaining state variables must be computed.
38
The other limit to be considered in HYGOV is the gate velocity limit. In order to calculate the
dynamic effects of this type of limit, the input variable VELM is given. This variable represents
the reciprocal of the time taken for the gates to move from fully open to fully close. Therefore
VELM can be seen as the growth rate of the gate position. Recalling that the derivative of a
position is in fact a velocity, we can use the relations ( 4.37 ) and ( 4.39 ) to determine the
maximum and minimum desired gate position, due to velocity limits.
( 4.37 )
Gate
opening:
( 4.38 )
( 4.39 )
Gate
closing:
( 4.40 )
Here,
is the numerical integration solution,
is the present value and
is the simulation
time-step.
When computing the integration solution,
is higher than
than
is compared with
and
it means that the gate is opening too fast. Otherwise, if
, the gate is closing too fast. In both cases,
. If
is smaller
must be limited with the
respective restriction value.
4.2.5. GAST model

From the block diagram depicted in Figure 3.8 result the representative differential state
equations of the GAST model.
( 4.41 )
( 4.42 )
( 4.43 )
with the limit constraint of the valve opening
39
( 4.44 )
The fuel flow is controlled by the low value gate, which selects the lowest value between the
outputs of the load-frequency control and the temperature control. These are respectively given
by the algebraic equations ( 4.45 ) and ( 4.46 ).
( 4.45 )
( 4.46 )
4.3. Models in Algebraic Form

The differential expressions presented in the last Section need to be converted into an algebraic
state equation so that the integration algorithm represented by ( 4.1 ) can be applied. The state
function has the form of the algebraic expression ( 4.47 ).
( 4.47 )
This outline takes into account that the time constants and the parameters associated with the
dynamic models remain constant throughout the simulation process and, therefore, do not need
to be computed in every time step.
Matrix
includes the dependent associated time constants and model parameters, constant in
all the computation.

Matrix
retrieves the non-constant terms, thus requiring to be calculated in every time step.
Matrix
includes the independent terms, related with the matrix
, which contains the fixed
inputs. These matrices are also constant throughout the simulation, although there might be
cases where some variables of
This way, only matrix
need to be calculated in every time step.
needs to be consistently computed.
This schema is very useful in a computational effort point of view, since it optimizes the program
preventing unnecessary calculations, and consequently increasing simulation speed.
The following Subsections give the algebraic state equations of the considered models. Each
dynamic model is exhibited separately, which is not a realist case since governor and excitation
control systems must be connected with a generator unit, thus constituting the generator group.
Nevertheless it was chosen to display the models like this, because the main concern of this
Chapter is in fact the isolated models themselves.
40
Appendix C provides four different combinations of synchronous generator + excitation system

+ speed governor system, in order to give an insight of the complete generator group
arrangement.
4.3.1. GENROU and GENROE models

The GENROU and GENROE dynamic model represented by the differential equations ( 4.4 ) (
4.9 ), in its representative algebraic form, is given by the state function ( 4.48 ).
( 4.48 )
The matrices
and
are defined by
( 4.49 )
41
( 4.50 )
( 4.51 )
( 4.52 )
4.3.2. IEEET1 model

The IEEET1 dynamic model represented by the differential equations ( 4.19 ) ( 4.22 ), in its
representative algebraic form, is given by the state function ( 4.53 )
( 4.53 )
The matrices
and
are defined by
42
( 4.54 )
( 4.55 )
( 4.56 )
( 4.57 )
4.3.3. IEEEX1 model

The IEEEX1 dynamic model represented by the differential equations ( 4.24 ) ( 4.28 ), in its
representative algebraic form, is given by the state function ( 4.58 ).
( 4.58 )
The matrices
and
are specified by
43
( 4.59 )
( 4.60 )
( 4.61 )
( 4.62 )
4.3.4. HYGOV model

The HYGOV dynamic model represented by the differential equations ( 4.30 ) ( 4.33 ) , in its
representative algebraic form, is given by the state function ( 4.63 ).
( 4.63 )
The matrices
and
are defined by
44
( 4.64 )
Matrix A of equation ( 4.64 ) is not complete. From equations ( 4.30 ) and ( 4.31 ) we can see
that the derivatives of both the filter output and the desired gate position depend on the speed
of the machine, which is a state variable of the synchronous generator. Since the dynamic
models presented in this Section are isolated, the state variables from other models cannot be
represented. To overcome this setback,
and its dependent terms are placed into matrix R.
However, it should be noted that this is merely a representation, because a speed governor is
always connected to a generator.
( 4.65 )
( 4.66 )
( 4.67 )
4.3.5. GAST model

The GAST dynamic model represented by the differential equations ( 4.41 ) ( 4.43 ), in its
representative algebraic form, is given by ( 4.68 )
( 4.68 )
The matrices
and
are defined by
45
( 4.69 )
( 4.70 )
( 4.71 )
( 4.72 )
4.4. Initial Conditions for Dynamic Simulation

This Section delivers the expressions of the initial conditions, for each one of the dynamic
models presented in the last Sections.
The correct computation of the initial values of the dynamic states is crucial for the accuracy of
the results. Also, the initial conditions provide an assessment point for the simulation. If these
values are not in agreement with those obtained by PSS/ETM, it means that there are preceding
errors. These may exist in the implemented dynamic model equations, in programming errors or
even in the initial conditions computation themselves. It should be noted that the load flow
results obtained in previous stages are the basis for the calculation of the initial values, as well
as for the fixed inputs.
4.4.1. GENROU and GENROE Initial Conditions

The computation of the generator initial values follows the suggested steps in Section 2.1.
1. Compute the generator current in the network frame
( 4.73 )
The generated active and reactive powers values, as the complex voltage, are obtained
from the load flow calculation.
2. Compute the initial value of the rotor angle,
) and substituting
. This is obtained by setting
in ( 4.6
in the algebraic equation ( 4.15 ) resulting
46
( 4.74 )
3. Transform the current and the voltage from the network reference frame to the
generator reference frame.
( 4.75 )
( 4.76 )
4. Calculate the transient emf of the direct axis,
. There are two ways to compute this
value.
One way is by setting
and
in equations ( 4.6 ) and ( 4.7 ) respectively. Then
is substituted in ( 4.6 ). After some algebraic manipulation,
is obtained.
( 4.77 )
The other way to compute this value is using ( 4.78 ) [8].

( 4.78 )
This serves as a check on the calculations.
Using
and setting
in ( 4.7 ),
is computed.
( 4.79 )
5. Compute the transient emf of the quadrature axis,

substituting
, by annulling equation ( 4.5 ) and
in the algebraic equation ( 4.15 ).
( 4.80 )
6. The initial value of the electrical field voltage is found by cancelling both ( 4.4 ) and
( 4.5 ) and subsequently substituting
in ( 4.4 ). The result is expressed in

( 4.81 )
7. The mechanical torque initial value is found by setting the derivative term of ( 4.9 ) equal
to zero, resulting the expression
47
( 4.82 )
For the computation of
it was taken into account that
and therefore
4.4.2. IEEET1 Initial Conditions

After determining the initial conditions of the generator (mainly,
) the initial values of the
state variables of the exciter can be found. This procedure is described below.
1.
is acquired by making
algebraic manoeuvring the
equal to zero in equation ( 4.19 ). Then, after the required

initial condition is obtained.
( 4.83 )
2. Rate feedback state variable,
is found by cancelling equation ( 4.20 ).

( 4.84 )
3. The sensed terminal voltage is acquired by setting
in ( 4.22 ).
( 4.85 )
4. Finally, after getting the values expressed in ( 4.83 ), ( 4.84 ), ( 4.85 ) the reference
voltage is found by setting
in ( 4.21 ) and by also performing some algebraic
manipulation.
( 4.86 )
4.4.3. IEEEX1 Initial Conditions

The exciter model IEEEX1 initial conditions computation is very similar to the one performed for
IEEET1. However, there are some differences due to the inclusion of the lead-lag block.
1.
is found by setting the
derivative equal to zero in equation ( 4.24 ). The result is
expressed in ( 4.87 ).
( 4.87 )
2. Rate feedback state variable,
is found with annulment equation ( 4.25 ).
48
( 4.88 )
3. The sensed terminal voltage is acquired by setting
in ( 4.27 ).
( 4.89 )
4. The lead-lag block output can be attained by setting
in equation ( 4.26 ), which
results in equation ( 4.90 ).

( 4.90 )
5. In this dynamic model the reference voltage is found by using the lead-lag block
differential equation ( 4.28 ), where it should be set
( 4.91 )
4.4.4. HYGOV Initial Conditions

After the computation of the initial value of the mechanical torque, the calculation of the initial
conditions of HYGOV is done.
1. By setting
in ( 4.33 ), the water head initial value is obtained

( 4.92 )
2. The water flow initial value is reached from the manipulation of the algebraic equation (
4.35 ), in which results
( 4.93 )
It should be noticed that
, since
3. The gate opening initial value is reached from the manipulation of the algebraic
equation ( 4.34 ) , in which results
( 4.94 )
4. By setting
in equation ( 4.32 ) the desired gate positioning is achieved.
49
( 4.95 )
5. In order to compute the two missing state variables,
and
, some manipulation is
required. First, the differential equations ( 4.30 ) and ( 4.31 ) are equalled to zero. Then
the following algebraic system is formed:
Solving this system, the initial values for the filter output and for the speed reference are found.
It should once again be observed that
.
( 4.96 )
( 4.97 )
4.4.5. GAST Initial Conditions

1.
The gas turbine governor mechanical power reference is immediately obtained after the
computation of the mechanical power initial value.
( 4.98 )
2.
By cancelling ( 4.42 ), the fuel valve opening initial position is obtained

( 4.99 )
3.
By setting
in ( 4.41 ), the fuel flow state variable is acquired

( 4.100 )
4.
Finally the exhaust temperature initial value is obtained by cancelling ( 4.42 )

( 4.101 )
50
5. Simulation Results
Contents
5.
Simulation Results ........................................................................................................... 51

5.1.
Dynamic Simulation Procedure ................................................................................ 52
5.2.
Validation of the Dynamic Models ........................................................................... 52
5.2.1.
GAST Validation ............................................................................................... 54
5.2.2.
IEEEX1 Validation ............................................................................................. 56
5.2.3.
IEEET1 Validation ............................................................................................. 59
5.2.4.
HYGOV Validation ............................................................................................ 60
5.3.
57-Bus Case ............................................................................................................. 64

51
This Chapter presents the performed simulations. At first, we make a time event narrative of
what happens in the dynamic simulation. The next step is the validation of the described models
in the previous sections. There is a set of four demonstrations, each one with an explanation of
the behaviour of the variables. The simulation results of the 57-Bus case are given in the last
stage of this Chapter. From the results of this simulation it is given a comparison between the
different dynamic models.
Although only five simulations are presented here, many more were done during the completion
of this work. However, these cases are the ones that better reveal the improvements made in
both the revisited and the newly implemented dynamic models. Appendix F also provides the
simulation results of a 30-Bus case. It was chosen to include this case in the appendices since it
served as a bridge between the smaller networks and the 57-Bus network, thus playing an
important role on the performed work.
Due to the lack of space, both the dynamic input data and the initial values of the state variables
of all the presented simulations are shown in Appendix C.
5.1. Dynamic Simulation Procedure

This software package simulates symmetric three phase short-circuits in a particular bus and
the consequent isolation of the fault by removing a line, in order to examine the response of the
power system and its integrated machines.
The simulation begins at
and, one second later, the short circuit is applied to a specific
branch near a bus. Until then, the system does not suffer any changes, maintaining its
stationary values, accordant to power flow results and the machine initial conditions. At
the fault is cleared by removing the faulted branch, thus enacting the opening of the
circuit breaker. In order to achieve steady state regime once again, the simulation is carried out
to
, time at which the oscillations of the power system should have disappeared
completely. If not the system may be unstable.

For all the simulations exhibited in this dissertation, the systems base is
frequency is
, the nominal
, and, with the exception in the HYGOV validation, the used time-step is
.
5.2. Validation of the Dynamic Models

This Section performs the validation of the models. There is a series of four simulations, which
include the following model combinations:
GENROU + IEEET1 + GAST
GENROE + IEEEX1 + TGOV1
GENROE + IEEET1 + TGOV1
GENSAL + IEEET1 + HYGOV
52
The first simulation is intended to validate the GAST model, while providing an explanation for
the response of both the system and the generator group. This description is valid for most of
the simulations carried out.
The second simulation validates the IEEEX1 dynamic model.
The third and fourth simulations offer an overview of the effects of the limit constraints that were
implemented in the synchronous generator control models.
The 2-Bus network presented in Figure 5.1 is used in these simulations. Table 5.1 gives the
results obtained by the load flow computation.
The symmetric three-phase short circuit is applied in Branch 2 near Bus 1. Branch 2 is
consequently tripped in order to replicate the opening of the protection system.
Figure 5.1 2-Bus network

Table 5.1 Power flow results for the 2-Bus simulation
Power Flow Results

BUS
1
2
Voltage
Swing
P-Q
1.0400 0.0000
1.0247 -0.5877
0.2508
-
0.0613
-
0.2500
0.5000
Due to the large number of variables that can be represented, it was necessary to make a
selection. For the newly implemented models, the illustrated variables are:
Bus voltages,
and
Machine terminal current,

Generated Active Power,
Generated Reactive Power,
Mechanical Power,
Rotor angle,
Speed deviation,
Regarding the IEEET1 validation, the variables related with the excitation system are displayed.
In the case of the HYGOV simulation, the variables which allow verifying the limit operation of
the governor are shown.
As stated in previous Chapters, the results of the simulations are compared side-by-side with
TM
the output of the recognized simulation software PSS/E . In all figures the results of the
53
MATLAB program are represented by a black continuous line and the PSS/ETM results are
represented by a yellow filled area.
5.2.1. GAST Validation

In the last Section it was given the 2-Bus network and the power flow results, used in the
simulation of the generator group, composed by GENROU + IEEET1 + GAST models. Figure
5.2 and Figure 5.3 depict the results of the simulation. The dynamic parameters of the generator
group and initial conditions are displayed in Appendix C.1, while the complete coupled model is
presented in Appendix D.1.
(a) Voltage magnitude, Bus 1
(c) Generator Active Power
(b) Voltage magnitude, Bus 2
(d) Generator Reactive Power
(e) Generator Terminal Current

Figure 5.2 GAST model validation results (Part 1 of 2)
54
(a) Generator Mechanical Power
(b) Speed deviation
(c) Generator Rotor Angle
(d) Exciter Field Voltage
Figure 5.3 GAST model validation results (Part 2 of 2)
Before
, that is, before the disturbance, the system is in a stationary regime, in conformity
with the results found in the power flow and initial conditions computation. When the threephase short circuit is applied, the voltage magnitudes of both buses instantaneously decay to
zero, as seen in Figure 5.2(a) and Figure 5.2(b), due to the low impedances caused by the short
circuit. Also, as a consequence, an extremely high valued short circuit current emerges. This
short circuit current is fed by the generator, as shown in Figure 5.2(e), where the terminal
current of the generator rises abruptly and maintains its elevated value during the fault. This
sudden growth of
is due to the fact that magnetic energy cannot be instantly cancelled.
Because of this, the magnetic flux in the air gap has to be preserved at its initial value after the
short circuit. As a consequence a high valued current emerges in the stator of the machine.
Figure 5.2(c) and Figure 5.2(d) show that active and reactive powers decrease, following the
tendency of the voltage magnitudes.
The instant the short circuit is applied, the governor and excitation control systems respond to
the variations of the power system. During the fault, active power decreases immensely and,
the load demand cannot be supplied. Because of this, a mismatch between the mechanical and
55
the electrical torques occurs, which results in an increase of the speed of the machine. This is
marked by Figure 5.3(b).The governor, becoming aware of this speed increase, acts on the
turbine valve, by closing it. This consequently results in a decrease of the mechanical power
(Figure 5.3(a)) in an attempt to approximate it to the electrical power. This way, as the two
powers come closer together, the machine speed is reduced.
Regarding the excitation system, it is visible in Figure 5.3(d), that when the fault is applied the
generator excitation voltage,
, raises its value, in order to compensate the weakening of the
air gap flux.

Comparing Figure 5.3(d) with Figure 5.3(a) it is perceptible the much faster response of the
excitation control system when compared with the governor control system. This was already
expected, as the excitation system has the smallest time constants of the group.
After 150ms the protective system eliminates the short circuit, and as Figure 5.2(e) shows, the
generator terminal current returns to a value close to that of the initial situation. Consequently
the bus voltages rise. Because of the time lags intrinsic to the voltage regulator, the voltage
magnitudes surpass the initial condition values, causing an overshoot. Once again active and
reactive powers follow the tendency of the generator voltage, also exceeding their initial values
with a rising swing. This peak of the active power, together with a decreased mechanical power,
reduce the speed of the machine which in its turn makes the governor to raise the mechanical
power, to once again try to eliminate the speed deviation.
Figure 5.2(d) also shows that the stationary value of the reactive power changes.
increases
in order to compensate the removal of the Branch 2, which was mainly inductive.
As time moves forward, a stationary regime is achieved, and a new network balance is found
with the adaptation of the system to the new topology of the grid.
5.2.2. IEEEX1 Validation

For the validation of the IEEEX1 model a new simulation was performed using once again the
2-Bus network shown in Figure 5.1, and the power flow results displayed in Table 5.1. The input
parameters of the dynamic models can be consulted in Appendix C.2, while Appendix D.2 gives
the complete algebraic state equation of the generator group.
Figure 5.4 and Figure 5.5 presents the results of the simulation.
56
(a) Voltage magnitude, Bus 1
(b) Voltage magnitude, Bus 2
(e) Generator Terminal Current
Figure 5.4 IEEEX1 model validation results (Part 1 of 2)
57
(a) Generator Mechanical Power
(b) Speed deviation
(d) Exciter Field Voltage
Figure 5.5 IEEEX1 model validation results (Part 2 of 2)
When the disturbance occurs, it is once again possible to observe a dip in both voltages (Figure
5.4(a) and Figure 5.4(b)), as it is in active and reactive powers (Figure 5.4(c) and Figure 5.4(d)).
After clearance of the fault, those variables start to rise in a damped manner. This is due to the
arrangement of the time constants of the lead-lag block. By changing these parameters, the
system sets totally different behaviours. In the presented simulation,
is big while
is very
small. A Root-Locus analysis would show that the arrangement of these time constants is in fact
5
responsible for the slow and damped response of the system . If we were to consider a big
and a small
, a new Root-Locus analysis would show that the system becomes unstable. If
the two time constants are in the same order of value, the response of the excitation system is
less damped and faster.
As it is seen in Figure 3.4,

the lead lag block
is responsible for the zero while
is responsible for the pole introduced in
58
5.2.3. IEEET1 Validation

This Section focuses on the IEEET1 model. Here, the limit restrictions of the voltage regulator
are analyzed. The results of this simulation are found in Figure 5.6. The dynamic data input and
the complete algebraic state equation of the generator group are respectively shown in
Appendix C.3 and Appendix D.3.
(a) Voltage Magnitude, Bus 1
(b) Voltage Magnitude, Bus 2
(f) Exciter Field Voltage

Figure 5.6 - IEEET1 model validation results
59
In order to verify the response of the excitation system, when the voltage regulator limits are
violated, the maximum limit of the regulator is set with a small value6. Hence, as a result of this
7
value, the infringement endures from almost the instant the fault is applied until the end of the
simulation, that is,
is equal to
for most of the simulation.
When the disturbance occurs and the bus voltage magnitudes decrease instantaneously, the
exciter system responds by increasing the signal to the voltage regulator, whose output is
eventually bigger than allowed. When the limit is reached, the voltage regulator output is
instantaneously set with the limit value in the form of a step function.
Equation ( 4.19 ) shows that
is the derivative of two terms. One term includes
the other contains the output variable of the regulator. As

function, and its value remains constant, the growth rate of
response of
. This growth stops when
itself and
is limited in the form of a step

will be smaller a filtered
reaches its stationary value, which is derived from
( 4.19 ) and given by
5.2.4. HYGOV Validation

In the description of the Hydro-Turbine governor presented in Section 4.2.4, two types of limit
restrictions of the governor's gate were introduced the gate position limit and the gate velocity
limit. This Section presents a simulation to prove the implementation of both of these limits. A
simple analysis of the behaviour of the governor is also given.
Figure 5.7 and Figure 5.8 presents the results of the simulation. The dynamic parameters are
given in Appendix C.4 and the complete coupled models are presented in Appendix D.4.
Special attention must be paid to this simulation case. As stated, this simulation intends to see
what happens when limit restrictions of HYGOV are broken. It was observed that, when the
position limits constraints were broken, the program produced inaccurate results. It was later
verified that the problem resided on the used time-step value of the simulation. In order to
obtain accurate results, the time-step has to be reduced. So, in this case,
is equal to
Although the problem resided in the MATLAB program, the simulation on PSS/E
with
TM
was also set
. This was made so that the simulations in both programs were performed in the
exact same conditions.
6
7
It is reminded that the excitation control system has very small time constants, which makes the model to
have a very fast response.
60
(a) Generator Active Power
(b) Generator Mechanical Power
(d) Speed Deviation
Figure 5.7 - HYGOV model validation results
61
(a) Desired gate position state variable

response for the entire simulation.
(b) Response after the fault gate
(c) Gate position limits transgression
velocity limits transgression

Figure 5.8 Desired gate position state variable response
When the short circuit occurs, the electrical active power, as already seen in the previous
sections, dips instantaneously to zero, increasing the difference between electrical and
mechanical torques. This, due to the swing equation, raises the speed of the machine. The
governor detects the rapid speed increase and orders the turbine to close its gates in order to
decrease the mechanical power transmitted to the shaft of the generator. However, the velocity
limits of the hydro-turbine gates are very small, and as soon the governor enters in action, these
limits are broken, restricting the gate position. This is visible in Figure 5.8(b), where the desired
gate position behaviour is plotted. When the fault occurs, this variable has a linear closing
response, revealing a constant velocity the limit gate velocity. This has repercussions on the
mechanical power. Combining equation ( 4.34 ) with ( 4.35 ) and disregarding the turbine
damping, we can see that the mechanical power is proportional to the position of the gate. By
consequence, this implies that the variation rate of the mechanical power is in fact proportional
the variation rate of the gate position.
After the clearance of the fault, the generated electrical power increases and, as a result, so
does the mechanical power. Due to the delays of the governor system, the gate only starts to
62
open at approximately
. Again in Figure 5.8(b) we can see that the gate opens at a
linear rate, meaning that the gate velocity limits were broken again.
At around
a new event occurs; the violation of the maximum boundary of the gate
position. As the mechanical power increases in order to catch up with the electrical power, the
gate reaches its maximum opening, and therefore, cannot open any further. This is observed in
Figure 5.8(c) where the desired gate position has a maximum value of
defined
, which is the
At approximately
, the electrical power reaches a steady state regime and the governor
finally has a fixed value to follow. Since at this time the generated power is smaller than the
mechanical power, the hydro governor closes its gates in order to reduce the difference of the
two quantities. As seen in Figure 5.8(a), no limits are broken at this time.
63
5.3. 57-Bus Case

This Section presents the results of the 57-Bus IEEE network simulation. The topology of the
network is represented in Figure 5.9, whereas Table 5.2 gives the results of the load flow
computation.
Figure 5.9 Single-line diagram of the 57-Bus network
64
Table 5.2 Power flow results for the 57-Bus simulation.

Power Flow Results
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
BUS
Swing
P-V
P-V
P-Q
P-Q
P-V
P-Q
P-V
P-V
P-Q
P-Q
P-V
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
P-Q
Voltage
1.0400 0.0000
1.0100 1.3644
0.9850 0.5522
0.9796 0.1897
0.9761 0.4927
0.9800 1.0991
0.9823 -0.9774
1.0050 0.0327
0.9800 -2.0483
0.9834 -4.1283
0.9707 -3.5070
1.0150 -2.8624
0.9781 -3.7933
0.9710 -3.7820
0.9852 -2.2859
1.0207 -2.4616
1.0212 -2.5692
0.9577 -7.9801
0.9250 -9.0583
0.9176 -8.9306
0.9154 -7.9254
0.9166 -7.7611
0.9150 -7.7937
0.9053 -7.4366
0.8334 -18.7511
0.9068 -7.0440
0.9370 -5.3290
0.9553 -4.1689
0.9710 -3.3830
0.8131 -19.2081
0.7889 -19.3464
0.8177 -16.9444
0.8150 -16.9980
0.8625 -9.6820
0.8722 -9.3056
0.8844 -8.9285
0.8936 -8.6356
0.9199 -7.6569
0.8921 -8.6797
0.8835 -8.9816
0.9298 -8.4914
0.8857 -10.1253
0.9578 -4.9534
0.9327 -6.9176
0.9713 -4.5545
0.9562 -5.6964
0.9332 -7.2639
0.9294 -7.3808
0.9363 -7.3509
0.9290 -7.4740
0.9731 -5.6697
0.9333 -4.9779
0.9200 -5.6526
0.9404 -4.7240
0.9707 -3.4209
0.8760 -10.6869
0.8667 -11.4109
0.8807
0.8000
1.0000
1.0000
1.5000
0.8000
1.5000
-
1.9467
-0.2421
-0.2287
-0.3101
0.9420
-0.2939
0.8869
-
0.1500
0.0300
0.2100
0.1300
0.3500
0.5000
0.7000
0.0500
1.0000
0.5800
0.1050
0.2200
0.1300
0.4200
0.2720
0.0330
0.0230
0.0630
0.0630
0.0930
0.0460
0.1700
0.0360
0.0580
0.0160
0.0380
0.0600
0.1400
0.0630
0.0710
0.0200
0.1200
0.2970
0.1800
0.2100
0.1800
0.0490
0.2000
0.0410
0.0680
0.0760
0.0670
0.1700
0.8800
0.2100
0.0400
0.0200
0.2200
0.2600
0.0200
0.2400
0.0230
0.0530
0.0500
0.0300
0.0800
0.0980
0.0060
0.0100
0.0210
0.0320
0.0050
0.0230
0.0260
0.0180
0.0290
0.0080
0.0190
0.0300
0.0700
0.0300
0.0440
0.0100
0.0180
0.1160
0.0850
0.1050
0.0530
0.0220
0.1000
0.0140
0.0340
0.0220
0.0200
65
This simulation also follows the procedures stated in Section 5.1. The three phase short circuit
is applied to the branch that connects Buses 42 and 57, near Bus 42,
As seen in Figure 5.9, this network has a set of seven generation stations to supply the
necessary power for the distributed demands across the whole system. Table 5.3 presents the
different combinations of the used dynamic models, as well as their location in the network.
Table 5.3 Dynamic models used in the 57-Bus simulation
Bus
Generator Group Combination
GENROE + IEEET1 + GAST
GENROE + IEEEX1 + TGOV1
GENSAE + IEEET1 + HYGOV
GENROU + IEEET1 + TGOV1
GENROU + IEEET1 + GAST
12
Due to the large dimensions of the case in study, the total number of figures to display is
enormous8, therefore only a selection of results is delivered. Regarding the network, the
voltages of buses 42 (bus near short circuit), 56, 41 (buses electrically-adjacent from the short
circuit), 1 and 2 (electrically-distant buses from the short circuit) are displayed. For each
generator group a set of six images is shown rotor angle, generated active and reactive
powers, mechanical power, generator electrical field voltage and, lastly, the speed deviation.
These results are depicted in Figure 5.10 to Figure 5.17. The dynamic input data and the initial
conditions are given in Appendix C.5. For this simulation case, the errors of all of the above
mentioned variables are also provided. These errors are shown in Appendix E.
In order to represent all the variables of interest, there were about one hundred images to be plotted.
66
(c) Voltage Magnitude, Bus 56
(d) Voltage Magnitude, Bus 1
(e) Voltage Magnitude, Bus 2
Figure 5.10 57-Bus simulation results Voltages
67
(d) Generator Rotor Angle
(b) Generator Reactive Power
(e) Generator Mechanical
(c) Exciter Field Voltage
(f) Speed Deviation
Power
Figure 5.11 57-Bus simulation results - Gen. group at bus 1, GENSAL + IEEET1 + HYGOV
(f) Speed Deviation
Power
Figure 5.12 57-Bus simulation results - Gen. group at bus 2, GENROE + IEEET1 + GAST
68
(f) Speed Deviation
Power
Figure 5.13 57-Bus simulation results - Gen. group at bus 3, GENROE + IEEEX1 + TGOV1
(f) Speed Deviation
Power
Figure 5.14 57-Bus simulation results Gen. group at bus 6, GENSAE + IEEET1 + HYGOV
69
(f) Speed Deviation
Power
Figure 5.15 57-Bus simulation results Gen. group at bus 8, GENROU + IEEET1 + TGOV1
(f) Speed Deviation
Power
Figure 5.16 57-Bus simulation results Gen. group at bus 9, GENROU + IEEET1 + GAST
70
(f) Speed Deviation
Power
Figure 5.17 - 57-Bus simulation results Gen. group at bus 12, GENSAL + IEEET1 + HYGOV
Through the observation of the figures, it can be seen that the results of the developed software
have the same dynamic behaviour as the one provided by the PSS/ETM results.
Regarding the voltages, Figure 5.10
shows that, when the fault is applied, they suffer a
decrease in their magnitude. The voltage in bus forty two dips to zero, since this is a bus
electrically-close to the fault. By observing Figure 5.10, it is also possible to detect that the
electrical distance is an important factor in the behaviour of the voltages. During the
disturbance, the voltages of the adjacent buses experience great reductions, while in buses
farther apart the decrease of the voltage magnitudes is small.
Concerning the governor systems, we can see the differences between the three implemented
systems. It is perceptible that the hydraulic-turbine governor is the slowest one, mostly due to
the time lags related with the hydraulic system. The gas turbine is the fastest one, once again
due to its time constants, which are smaller than the ones of the steam and hydro turbines.
Another observed distinction between HYGOV and the other two models is the slight initial
overshoot verified in the mechanical power of the hydro-turbine governor, which is a typical
behaviour of these kinds of turbines.
Regarding the excitation systems, we can see that their responses are all quite similar, with the
exception of the generators of buses two and three. In the case of the generator group of bus
two, the excitation system IEEET1 is considering the delay caused by the terminal voltage
transducer. This delay makes the excitation system to respond slower when compared with
71
other systems that do not consider this time lag. The generator group in bus three uses the
excitation system IEEEX1 (or Type DC1A). As shown in Chapters 3 and 4, this system
considers additional time delays in the voltage regulator, represented by the time constants of
the lead-lag block. These time constants are, therefore, responsible for the slower response of
IEEEX1.
A close examination of the speed deviation figures shows that the results of the MATLAB
TM
program are not in total agreement with the ones obtained by PSS/E . In all the generators, the
computation of the first swing of all the represented variables is correct and concordant with
PSS/ETM. However, as the simulation advances in time,
suffers a shift, experiencing a
TM
delayed response when compared with PSS/E . As a consequence, electrical and mechanical
powers also suffer some deviations, resulting in small mismatches between the two simulation
packages.
The errors of all the represented variables in this simulation case can be consulted in Appendix
E. The error of speed deviation is not included because it often bounces from positive to
negative values crossing the zero line very often. Due to the previously mentioned speed shift,
every time the PSS/ETM plot has a null value, the error percentage rises to infinite values,
preventing a clear analysis.
72
6. Conclusions
Summary of the Work
This work had the goal of improving a computer program used for dynamic simulations
dedicated to transient stability studies in electrical power systems. This software, which has
been developed over the last years by IST students, was already equipped with several
dynamic models and, until the completion of this work, offered the possibility of simulating
networks up to nine buses.
The work focused on three major upgrades:
1. to pick up the program in the point where it was left at and give it the ability to deal with
larger networks;
2. to upgrade the existing dynamic models, providing them with features that, until the work
performed here, had not been implemented;
3. to increase the variety of choice given by the simulation software's library, with the addition
of new dynamic models.
The first part of this work consisted in learning the routines and the procedures of the program,
in order to detect possible flaws and to find the reasons that hindered the simulation of larger
networks. This was the most time consuming phase due to the extensive contents that reside in
a program of this type; data acquisition, the computation of both the power flow and the
dynamic states and, finally the presentation of the results. During this stage, the solutions that
enabled the program to deal with realistic networks were implemented. Chapter 2 presents a
description of the program, as well as of some of the improvements made to its structure.
Therefore, this Chapter can be used as a manual of the simulation software.
In the following stage, the work focused on the dynamic models of the software's library. When
testing these models in larger networks, it was found that they had some limitations. Thus, the
existing models were studied in depth in order to detect its problems and to solve them. In one
way or another, all the models suffered changes, being that the most important were made in
the control systems of the generators. The most significant limitation of these systems was that
the limits of both governor and excitation control systems were not yet taken into consideration
by the software. In governing systems, these limits represent the physical restrictions in the
valves/gates of the turbines. The limits in excitation systems are imposed so that the voltage
regulator does not produce outputs that exceed practical limits. In case of boundaries violation
73
of the control systems, it was necessary to implement an iterative solution within the numerical
calculations, in order to compute the response of all the generator group variables. There was
another important upgrade made to the IEEE Type I model (or IEEET1). This excitation system
failed to consider the terminal voltage transducer system. This was implemented in order to
account for the additional delays caused by this voltage measurement system.
Thereafter, two new dynamic models were designed and implemented the turbine governor
system GAST, which consists on a gas turbine, and the excitation control system Type DC1A
(or IEEEX1), which is very similar to IEEE Type 1, but considers a more complete voltage
regulator with additional blocks.
Chapter 3 presents the majority of the studied models, while giving an insight on the importance
of the generator control systems. This Chapter provides a theoretical description of the
operating models and their dynamic blocks, exposing also the meaning and purpose of all the
parameters and time-constants of the dynamic models.
Chapter 4 presents the numerical solutions for the dynamic simulation. Initially, a new
integration method is presented the Modified Euler-Cauchy integration algorithm , in addition
to the already existing Fourth-Order Runge-Kutta integration algorithm. This new integration
method was included to grant the user the possibility of choosing an integration method closer
to that of PSS/ETM, which serves as a proof of the results obtained throughout the simulations.
Afterwards, the differential state equations and the algebraic expressions, which represent the
dynamics of each model, are presented. Subsequently, the representative differential equations
are converted to its algebraic form. This conversion enables the implementation of the dynamic
models in the simulation software. This Chapter also computes the expressions of the initial
conditions for each one of the represented models.
Chapter 5 presents the simulation results. At first, the dynamic simulation procedures are
described. The validation of the new and upgraded models is then presented. The credited
models are the newly implemented GAST and IEEEX1, and the improved IEEET1 and HYGOV.
These validations are composed by a set of four simulation cases, each one relative to a
specific dynamic model, involving a simple 2-Bus network. Besides recognizing the accuracy of
the implemented models, this Section also analyzes the behaviour of the power system. Special
attention is given to the response of the control systems, which were subjected to extreme
situations regarding their limit constraints. Through the observation of the results, it is possible
to conclude that all the models are correctly implemented, as their dynamic behaviour is similar
to the one obtained by PSS/ETM. However, it should be noted that, when the HYGOV governor
limits are broken, there is a necessity to use smaller time-steps in order to compute the dynamic
solutions accurately. This denotes numerical limitations when turbine-governors systems limits
are breached.
74
The simulation results of the 57-Bus IEEE network are presented in the end of Chapter 5. This
network comprises seven generators and forty-two loads. In this simulation case, it is also
possible to observe the similarities between the dynamic behaviours of MATLAB and PSS/E
TM
results. This is especially noticeable in the first swing and in the end of the simulation, when a
new steady state has been achieved. However, this simulation shows that the differences
between the results of the two software packages increase due to both the restricting action of
the governors and the growth of the network. This is mainly visible in the speed deviation,
which, after a certain point, begins to exhibit delays in comparison to PSS/E TM. Nevertheless, by
observing the percentage of the errors in this case, we verify that the differences between the
two programs always fall below 4%. This complies with the plotted objectives: the
implementation of a simulation software capable of dealing with large networks, while
accomplishing a level of precision very close to other simulation packages with a commercial
nature.
Future Work
At the end of this dissertation, the simulation software was left with a dynamic library consisting
of nine models, and with the ability to deal with cases up to fifty-seven buses, but nevertheless
with the required conditions to simulate even larger networks.
However, several improvements can still be made in the program. One of the main obstacles in
implementing new models is the intrinsic structure of the program, which is not generic. For
each generator group combination synchronous generator + turbine governor + excitation
system it is necessary to create a new specific case. As four different synchronous generators
are combined with the various control systems the number of combination cases is enormous.
So every time we want to add a new dynamic model, several new cases have to be created by
hand, in order that this new model is considered by all the generators. This is obviously an
exhausting and time consuming task. Therefore, it is important to perform a profound change in
the structure of the program so that the combination of the models is generic and transparent to
the user. Probably the best way to accomplish this is by employing a method of array indexation
and pointers adopted by PSS/ETM. More about this topic can be found in Section 11.4 of [11].
Another upgrade that may be done to the program is to give it the capability of having more than
one generator per bus, something that is still not possible to accommodate. While this is not an
arduous task, it requires some programming skills. The implementation of this feature may take
the program to other levels, as even more complex networks may be simulated.
Regarding the dynamic model library of the simulation package, this should be increased. A
suggested direction is the modelling of wind powered generation. The WTG class of PSS/ETM is
a good starting point for implementing this capability.
75
Although the results presented in this work are close to those obtained by PSS/E TM, there are
still some slight numerical differences, which can be further analyzed in order to achieve more
accurate results. However, it is necessary to note that, despite the fact that PSS/E
TM
is the
reference for the program in development, different calculation methods are used in both
programs, which therefore cause mismatches that will always persist. It is important to have a
critical view of the work and not to chase the unattainable.
76
Bibliography
[1]
I. P. W. G. on Prime Mover and E. S. M. for System Dynamic Performance Studies,

Hydraulic turbine and turbine control models for system dynamic studies, IEEE
Transactions on Power Systems, vol. 7, Feb. 1992.
[2]
IEEE Committee Report. "Computer representation of excitation systems." IEEE

Trans.Power Appar. Syst.,vol. PAS-87, no. 6. Jun. 1968.
[3]
IEEE Committe Report. "Generator reactive power limits and voltage stability." Power
System Control and Management, Fourth International Conference. Calgary, AB , Apr
1996.
[4]
IEEE Committee Report. "Gas Turbine Control for Islanding Operation of Distribution
Systems." IEEE Power & Energy Society General Meeting. Jul. 2009.
[5]
IEEE Std 421.5-2005. "IEEE Recommended Practice for Excitation System Models for
Power System Stability Studies." 2005.
[6]
Kundur, P. Power System Stability and Control. New York: McGraw-Hill, 1994.
[7]
P. Centeno, I. Egido, C. Domingo, F. Fernandez, L. Rouco, and M. Gonzalez. "Review of

gas turbine models for power system stability studies." 9th Spanish Portuguese Congress
on Electrical Engineering. Marbella, Spain, 2005.
[8]
Pai, P. W. Sauer and M. A. Power System Dynamics and Stability. Upper Saddle River,
NJ: Prentice-Hall, 1998.
[9]
Paiva, J. P. S. "Redes de Energia Eltrica: uma anlise sistmica". Lisboa, Portugal: IST
Press.
[10]
Paulo, Andr S.M. A Library of Dynamic Models for Transient Stability Studies. MSc
dissertation. Lisboa: Intituto Superior Tcnico, Universidade Tcnica de Lisboa, 2009.
[11]
PSS/E. "Dynamic Simulation Principles." Program Aplication Guide: Volume II.
[12]
PSS/E. "Excitation System and Controller Models." Program Application Guide: Volume
II.
[13]
PSS/E. "Excitation System Model Data Sheets." Program Operation Manual: Volume II.
77
[14]
PSS/E. "Explicit and Implicit Integration Algorithms." Program Application Guide: Volume
II.
[15]
PSS/E. "Generator Modeling." Program Application Guide: Volume II.
[16]
PSS/E. "Speed Governor System Modeling." Program Application Guide: Volume II.
[17]
PSS/E. "Turbine-Governor Model Data Sheets." Program Operation Manual: Volume II.
[18]
T.S. Parker, L.O. Chua. Practical Numerical Algorithms for Chaotic Systems. New York,
USA: Springler-Verlag, 1989.
78
Appendices
Contents
Appendix A - Windup and Non Windup limits.......................................................................... 80
Appendix B - Exciter Saturation Function................................................................................. 82
Appendix C - Dynamic Parameters and Initial Conditions for Dynamic Simulations .................. 83
C.1.
GAST Validation ....................................................................................................... 83
C.2.
IEEEX1 Validation ..................................................................................................... 84
C.3.
IEEET1 Validation ..................................................................................................... 84
C.4.
HGOV1 Validation .................................................................................................... 85
C.5.
57-Bus Network Simulation ..................................................................................... 86
Appendix D Algebraic State Equations for Validation Simulations ......................................... 89

D.1.
GAST Validation, GENROU + IEEET1 + GAST.............................................................. 89
D.2.
IEEEX1 Validation, GENROE+ IEEEX1 + TGOV1 .......................................................... 92
D.3.
IEEET1 Validation, GENROE+IEEET1+TGOV1 ............................................................. 95
D.4.
HYGOV Validation, GENSAL + IEEET1 + HYGOV ......................................................... 99
Appendix E 57-Bus Simulation Errors .................................................................................. 103

Appendix F 30-Bus Case ..................................................................................................... 111
F.1.
Dynamic Parameters.............................................................................................. 111
F.2.
Initial Conditions.................................................................................................... 112
F.3.
Simulation Results ................................................................................................. 113
79
Appendix A - Windup and Non Windup limits

The models described on this dissertation may differentiate between windup and non-windup
limits. Such limits can be encountered in integration blocks, single time constant blocks and
lead-lag blocks. Here, it is illustrated both type of limits for singe time constant blocks. The block
representations and the functions of these limits are presented in Figure A.1 and Figure A.2.
Figure A.1 Single time constant block - Windup limiter.
System Equation:
( A.1 )
If
Limiting Action:
, then
If
, then
If
, then
( A.2 )
Figure A.2 Single time constant block Non-Windup limiter.
System Equation:
( A.3 )
80
If
Limiting Action:
, then
If
, then set
If
, then set
( A.4 )
In the case of windup limits, if Y is outside of the limit range, the output voltage X is limited.
Therefore variable X can only come off the limit if Y has already entered the limit range.
In the case of non-windup limits variable Y is controlled. In order for Y to be limited, that is,
or
it is necessary that
or
respectively. Hence, Y comes off the
restriction range the instant u comes within the limits.
81
Appendix B - Exciter Saturation Function

An example of a saturation function is illustrated in Figure B.1
Figure B.1 Saturation Function (Source: [12])
This function is a design characteristic of the exciter and represents the increase in the exciter
excitation requirements due to saturation. In the simulation data input, in order to define the
saturation curve, two points of the curve are provided,
and
. From
these points, it will be able to compute the saturation factor value [12] using:
( B.1)
where A and B are defined as the exciter excitation required to produce that output voltage on
the constant-resistance-load saturation curve, and on the air-gap line, both represented in Figure
B.2.
Figure B.2 - Exciter saturation characteristics (Source: [5])
A and B are computed in the initial conditions are kept constant throughout the simulation.
82
Appendix C - Dynamic Parameters and Initial

Conditions for Dynamic Simulations
Here the parameters of the models and the initial conditions for the simulations presented in
Sections 5.2 and 5.3 are shown. The parameters represent the input dynamic files, and the
initial conditions are obtained when using the program.
C.1. GAST Validation

The dynamic parameters and initial condition of the simulation discussed in Subsection 5.2.1
are presented in the next subsections. In the simulation software this test is regarded as case
38.
C.1.1.
Dynamic Parameters
The parameters of the GENROU+IEEET1+GAST models that are used in the input *.dyr file are
presented below.
1 'GENROU'
4.00000
0.80000
1 'IEEET1'
-99.9999
0.00000
1 'GAST'
1.00000
C.1.2.
6.50000
0.00000
0.30000
0.00000
1.00000
0.00000
0.50000E-01
2.00000
0.60000E-01
1.80000
0.15000
200.000
0.30000
0.00000
0.40000
1.00000
0.20000
1.75000
0.00000
0.84000
0.67000E-01
0.00000
0.10000
-0.5000E-01
0.50000E-01
0.60000
0.00000
/
99.9999
1.00000
0.00000
/
3.00000
0.00000
/
Initial Conditions
The obtained initial conditions in the simulation for the GAST validation are:
Initial Conditions - GENROU BUS 1
83
C.2. IEEEX1 Validation

are presented in the next subsections. In the simulation software this test is regarded as case
36.
C.2.1.
Dynamic Parameters
The *.dyr file that contains the parameters for the GENROE+IEEEX1+TGOV1 simulations is
shown next:
1 'GENROE' 1
4.00000
0.80000
1 'IEEEX1' 1
1.000000
0.08000
0.00000
1 'TGOV1' 1
1.00000
C.2.2.
6.50000
0.60000E-01
0.00000
1.80000
0.30000
0.15000
0.08000
40.0000
4.00000
-4.0000
2.00000
0.00000
0.00000
/
0.50000E-01 0.50000
1.00000
0.00000
/
0.20000
1.75000
0.00000
0.20000
1.00000
0.00000
0.50000E-01
0.60000
0.00000
/
0.10000
0.50000
0.00000
1.00000
0.00000
Initial Conditions
The obtained initial conditions in the simulation for the IEEEX1 validation are:
Initial Conditions - GENROE BUS 1
C.3. IEEET1 Validation

are presented next. In the simulation software this test is regarded as case 34.
84
C.3.1.
Dynamic Parameters
The parameters of the GENROE+IEEET1+TGOV1 models that are used in the input *.dyr file
are presented below.
1 'GENROE'
4.00000
0.80000
1 'IEEET1'
-1.05000
0.00000
1 'TGOV1'
1.00000
C.3.2.
6.50000
0.00000
0.30000
0.10000
1.00000
0.00000
0.50000E-01
1.00000
0.60000E-01
1.80000
0.15000
200.000
0.30000
0.00000
0.50000
0.00000
/
0.20000
1.75000
0.00000
0.84000
0.67000E-01
0.00000
1.00000
0.50000E-01
0.60000
0.00000
/
1.55000
1.00000
0.00000
/
0.00000
Initial Conditions
The obtained initial conditions in the simulation for the IEEET1 validation are:
C.4. HGOV1 Validation

are presented next. This test is regarded as case 37. In this simulation, the time step is
.
C.4.1.
Dynamic Parameters
The parameters of the GENSAL+IEEET1+HYGOV models that are used in the input *.dyr file
are presented below.
1 'GENSAL' 1
0.000000
0.150000
1 'IEEET1' 1
-99.9999
0.000000
1 'HYGOV'
1
0.20000
2.50000
5.00000
1.50000
0.00000
0.00000
1.00000
0.00000
0.06000
0.01000
0.00000
0.40000E-01
0.90000
0.00000
/
200.000
0.30000
0.00000
0.40000
0.60100
0.50000
/
0.12000
0.60000
6.00000
0.30000
0.84000
0.06700
0.00000
8.00000
0.00000
99.9999
1.00000
0.00000
0.05000
1.20000
85
C.4.2.
Initial Conditions
The obtained initial conditions in the simulation for the HIGOV validation are:
Initial Conditions - GENSAL BUS 1
C.5. 57-Bus Network Simulation

The dynamic parameters and initial condition of the simulation discussed in Section 0 are
presented next. In the simulation software this test is regarded as case 603.
C.5.1.
Dynamic Parameters
The parameters of each one of the models that are used in the input *.dyr file are presented
next.
1 'GENSAL'
0.000000
0.120000
1 'IEEET1'
-4.00000
0.000000
1 'HYGOV'
0.500000
3.20000
2 'GENROE'
3.00000
0.60000
2 'IEEET1'
-2.00000
0.00000
2 'GAST'
1.00000
3 'GENROE'
3.00000
0.60000
3 'IEEEX1'
0.200000
0.08000
0.00000
3 'TGOV1'
0.40000
6 'GENSAE'
1
1
1
1
5.00000
0.50000E-01
1.50000
1.20000
0.00000
0.00000
/
0.00000
200.000
1.00000
0.30000
0.00000
0.00000
0.60000E-01 0.30000
0.20000
1.00000
0.00000
0.08000
/
6.50000
0.60000E-01
0.00000
1.40000
0.20000
0.10000
0.10000
200.000
1.00000
0.30000
0.00000
0.00000
0.50000E-01 0.40000
10.0000
1.00000
6.50000
0.60000E-01
0.00000
1.40000
0.20000
0.10000
0.08000
40.0000
2.00000
-2.0000
2.00000
0.00000
0.00000
/
0.50000E-01 0.40000
1.00000
0.00000
/
5.00000
0.50000E-01
0.06000
0.40000
5.08400
0.20000
0.84000
0.20000
0.00000
6.0000
0.00000
5.00000
1.00000
0.00000
0.05000
1.50000
1.00000
1.35000
0.00000
0.84000
0.67000E-01
0.00000
0.10000
-0.5000E-01
1.00000
1.35000
0.00000
0.84000
1.00000
0.00000
0.50000E-01
0.30000
0.00000
/
2.00000
1.00000
0.00000
/
3.00000
0.00000
/
0.50000E-01
0.30000
0.00000
/
0.10000
0.50000
0.00000
1.02000
0.00000
0.06000
5.08400
86
8
9
9
12
12
12
0.000000
0.120000
'IEEET1' 1
-7.24000
0.000000
'HYGOV'
1
0.500000
3.20000
'GENROU' 1
3.00000
0.60000
'IEEET1' 1
-4.50000
0.00000
'TGOV1' 1
0.40000
'GENROU' 1
3.00000
0.60000
'IEEET1' 1
-7.24000
0.000000
'GAST'
1
1.00000
'GENSAL' 1
0.000000
0.120000
'IEEET1' 1
-7.24000
0.000000
'HYGOV'
1
0.500000
4.20000
C.5.2.
1.50000
0.00000
0.00000
1.00000
0.00000
0.60000E-01
0.20000
0.00000
6.50000
0.00000
0.20000
0.00000
1.00000
0.00000
0.50000E-01
1.00000
6.50000
0.00000
0.20000
0.00000
1.00000
0.00000
0.50000E-01
10.0000
5.00000
1.50000
0.00000
0.00000
1.00000
0.00000
0.60000E-01
0.20000
0.00000
1.20000
0.00000
/
200.000
0.30000
0.00000
0.30000
1.00000
0.08000
/
0.60000E-01
1.40000
0.10000
200.000
0.30000
0.00000
0.40000
0.00000
/
0.60000E-01
1.40000
0.10000
200.000
0.30000
0.00000
0.40000
1.00000
0.50000E-01
1.20000
0.00000
/
200.000
0.30000
0.00000
0.30000
1.00000
0.08000
/
0.40000
0.20000
0.84000
0.20000
0.00000
6.0000
0.00000
7.24000
1.00000
0.00000
0.05000
1.50000
1.00000
1.35000
0.00000
0.84000
0.67000E-01
0.00000
1.52000
0.50000E-01
0.30000
0.00000
/
4.50000
1.00000
0.00000
/
0.00000
1.00000
1.35000
0.00000
0.84000
0.20000
0.00000
0.10000
-0.5000E-01
0.06000
0.40000
0.50000E-01
0.30000
0.00000
/
7.24000
1.00000
0.00000
/
3.00000
0.00000
/
5.08400
0.20000
0.84000
0.20000
0.00000
6.0000
0.00000
5.00000
1.00000
0.00000
0.05000
1.50000
Initial Conditions
--------------------------------------------------------------
--------------------------------------------------------------
87
Initial Conditions - GENSAE BUS 6
--------------------------------------------------------------
--------------------------------------------------------------
-----------------------------------------------------------------------------
--------------------------------------------------------------
--------------------------------------------------------------
88
Appendix D Algebraic State Equations for

Validation Simulations
This Appendix presents the complete machine matrices used in the algebraic state equations of
the simulations given in Chapter 5. Due to the large number of elements of these matrices, only
the non-zero terms are shown
D.1. GAST Validation, GENROU + IEEET1 + GAST

The algebraic state equation for the case simulated in Subsection 5.2.1 is:
( D.1 )
[A] matrix:
( D.2 )
( D.3 )
( D.4 )
( D.5 )
89
( D.6 )
( D.7 )
( D.8 )
( D.9 )
( D.10 )
( D.11 )
( D.12 )
( D.13 )
( D.14 )
( D.15 )
( D.16 )
( D.17 )
( D.18 )
( D.19 )
( D.20 )
( D.21 )
( D.22 )
90
( D.23 )
( D.24 )
( D.25 )
( D.26 )
( D.27 )
[R] matrix:
( D.28 )
( D.29 )
( D.30 )
( D.31 )
( D.32 )
( D.33 )
[C] matrix:
( D.34 )
( D.35 )
( D.36 )
91
( D.37 )
[u] matrix:
( D.38 )
( D.39 )
( D.40 )
D.2. IEEEX1 Validation, GENROE+ IEEEX1 + TGOV1

( D.41 )
[A] matrix:
( D.42 )
( D.43 )
( D.44 )
( D.45 )
92
( D.46 )
( D.47 )
( D.48 )
( D.49 )
( D.50 )
( D.51 )
( D.52 )
( D.53 )
( D.54 )
( D.55 )
( D.56 )
( D.57 )
( D.58 )
( D.59 )
( D.60 )
( D.61 )
93
( D.62 )
( D.63 )
( D.64 )
( D.65 )
( D.66 )
( D.67 )
( D.68 )
( D.69 )
( D.70 )
[R] matrix:
( D.71 )
( D.72 )
( D.73 )
( D.74 )
( D.75 )
94
( D.76 )
( D.77 )
[C] matrix:
( D.78 )
( D.79 )
( D.80 )
( D.81 )
( D.82 )
( D.83 )
( D.84 )
[u] matrix:
( D.85 )
( D.86 )
( D.87 )
D.3. IEEET1 Validation, GENROE+IEEET1+TGOV1

95
( D.88 )
[A] matrix:
( D.89 )
( D.90 )
( D.91 )
( D.92 )
( D.93 )
( D.94 )
( D.95 )
( D.96 )
( D.97 )
96
( D.98 )
( D.99 )
( D.100 )
( D.101 )
( D.102 )
( D.103 )
( D.104 )
( D.105 )
( D.106 )
( D.107 )
( D.108 )
( D.109 )
( D.110 )
( D.111 )
( D.112 )
( D.113 )
( D.114 )
[R] matrix:
97
( D.115 )
( D.116 )
( D.117 )
( D.118 )
( D.119 )
( D.120 )
[C] matrix:
( D.121 )
( D.122 )
( D.123 )
( D.124 )
( D.125 )
( D.126 )
( D.127 )
[u] matrix:
( D.128 )
( D.129 )
98
( D.130 )
D.4. HYGOV Validation, GENSAL + IEEET1 + HYGOV

( D.131 )
[A] matrix:
( D.132 )
( D.133 )
( D.134 )
( D.135 )
( D.136 )
( D.137 )
( D.138 )
99
( D.139 )
( D.140 )
( D.141 )
( D.142 )
( D.143 )
( D.144 )
( D.145 )
( D.146 )
( D.147 )
( D.148 )
( D.149 )
( D.150 )
( D.151 )
( D.152 )
( D.153 )
( D.154 )
( D.155 )
100
( D.156 )
( D.157 )
[R] matrix:
( D.158 )
( D.159 )
( D.160 )
( D.161 )
( D.162 )
( D.163 )
[C] matrix:
( D.164 )
( D.165 )
( D.166 )
( D.167 )
( D.168 )
( D.169 )
101
( D.170 )
( D.171 )
[u] matrix:
( D.172 )
( D.173 )
( D.174 )
( D.175 )
102
Appendix E 57-Bus Simulation Errors
(a) Voltage Magnitude error, Bus 42
Figure E.1 57-Bus simulation error percentages Voltages
103
(a) Generator Active Power error
(a) Generator Reactive Power error
(e) Generator Mechanical Power error
Figure E.2 57-Bus simulation error percentages - Gen. group at bus 1, GENSAL + IEEET1 + HYGOV
104
Figure E.3 57-Bus simulation error percentages - Gen. group at bus 2, GENROE + IEEET1 + GAST
105
Figure E.4 57-Bus simulation error percentages - Gen. group at bus 3, GENROE + IEEEX1 + TGOV1
106
Figure E.5 57-Bus simulation error percentages - Gen. group at bus 6, GENSAE + IEEET1 + HYGOV
107
Figure E.6 57-Bus simulation error percentages - Gen. group at bus 8, GENROU + IEEET1 + TGOV1
108
Figure E.7 57-Bus simulation error percentages - Gen. group at bus 9, GENROU + IEEET1 + GAST
109
Figure E.8 57-Bus simulation error percentages - Gen. group at bus 12, GENSAL + IEEET1 + HYGOV
110
Appendix F 30-Bus Case

This Appendix presents the results of a thirty bus case simulation. In the simulation software
this test is regarded as case 814.The short circuit is applied in bus four, and the removed
branch is the one that connects bus four and bus two.
F.1. Dynamic Parameters

1 'GENSAL'
0.000000
0.120000
1 'IEEET1'
-7.24000
0.000000
1 'HYGOV'
0.500000
1.20000
2 'GENROE'
3.00000
0.60000
2 'IEEET1'
-99.9999
0.00000
2 'GAST'
1.00000
5 'GENROE'
3.00000
0.60000
5 'IEEEX1'
1.000000
0.08000
0.00000
5 'TGOV1'
0.40000
8 'GENROU'
3.00000
0.60000
8 'IEEET1'
-99.9999
0.00000
8 'TGOV1'
0.40000
11 'GENSAE'
0.000000
0.120000
11 'IEEET1'
-99.9999
0.000000
11 'HYGOV'
0.500000
1.20000
13 'GENSAL'
0.000000
0.120000
13 'IEEET1'
-7.24000
0.000000
13 'HYGOV'
0.500000
1.20000
1
1
1
1
1
1
5.00000
0.50000E-01
1.50000
1.20000
0.00000
0.00000
/
0.00000
200.000
1.00000
0.30000
0.00000
0.00000
0.60000E-01 0.30000
0.20000
1.00000
0.00000
0.08000
/
6.50000
0.60000E-01
1.00000
1.40000
0.20000
0.10000
0.10000
200.000
1.00000
0.30000
0.00000
0.00000
0.50000E-01 0.40000
2.00000
1.00000
6.50000
0.60000E-01
1.00000
1.40000
0.20000
0.10000
0.08000
40.0000
99.9999
-99.999
2.00000
0.00000
0.00000
/
0.50000E-01 0.40000
1.00000
0.00000
/
6.50000
0.60000E-01
1.00000
1.40000
0.20000
0.10000
0.00000
200.000
1.00000
0.30000
0.00000
0.00000
0.50000E-01 0.40000
1.00000
0.00000
/
5.00000
0.50000E-01
1.50000
1.20000
0.00000
0.00000
/
0.00000
200.000
1.00000
0.30000
0.00000
0.00000
0.60000E-01 0.30000
0.16700
1.00000
0.00000
0.08000
/
5.00000
0.50000E-01
1.50000
1.20000
0.00000
0.00000
/
0.00000
200.000
1.00000
0.30000
0.00000
0.00000
0.60000E-01 0.30000
0.20000
1.00000
0.00000
0.08000
/
0.06000
0.40000
5.08400
0.20000
0.84000
0.20000
0.00000
6.0000
0.00000
7.24000
1.00000
0.00000
0.05000
1.50000
1.00000
1.35000
0.00000
0.84000
0.67000E-01
0.00000
0.10000
-0.5000E-01
1.00000
1.35000
0.00000
0.84000
1.00000
0.00000
0.50000E-01
0.30000
0.00000
/
99.9999
1.00000
0.00000
/
3.00000
0.00000
/
0.50000E-01
0.30000
0.00000
/
1.00000
0.50000
0.00000
0.45000
0.00000
1.00000
1.35000
0.00000
0.84000
0.67000E-01
0.00000
0.45000
0.50000E-01
0.30000
0.00000
/
99.9999
1.00000
0.00000
/
0.00000
0.06000
0.40000
5.08400
0.20000
0.84000
0.20000
0.00000
6.0000
0.00000
99.9999
1.00000
0.00000
0.05000
1.50000
0.06000
0.40000
5.08400
0.20000
0.84000
0.20000
0.00000
6.0000
0.00000
7.24000
1.00000
0.00000
0.05000
1.50000
111
F.2. Initial Conditions
-------------------------------------------------------------
--------------------------------------------------------------
--------------------------------------------------------------
--------------------------------------------------------------
112
F.3. Simulation Results
(c) Voltage Magnitude, Bus 3
(d) Voltage Magnitude, Bus 4
(e) Voltage Magnitude, Bus 5
(f) Voltage Magnitude, Bus 6
Figure F.1 30-Bus simulation results Voltage magnitudes

Power
(f) Speed Deviation
Power
Figure F.2 30-Bus simulation results Gen. group at bus 1, GENSAL + IEEET1 + HYGOV
113
(f) Speed Deviation
Power
Figure F.3 30-Bus simulation results Gen. group at bus 2, GENROE + IEEET1 + GAST
(f) Speed Deviation
Power
Figure F.4 30-Bus simulation results Gen. group at bus 5, GENROE + IEEEX1 + TGOV1
114
(f) Speed Deviation
Power
Figure F.5 30-Bus simulation results Gen. group at bus 8, GENROU + IEEET1 + TGOV1
(f) Speed Deviation
Power
Figure F.6 30-Bus simulation results Gen. group at bus 11, GENSAE + IEEET1 + HYGOV
115
(f) Speed Deviation
Power
Figure F.7 30-Bus simulation results Gen. group at bus 13, GENSAL + IEEET1 + HYGOV
116

HYGOV

Uploaded by

Copyright:

Available Formats

HYGOV

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

HYGOV

Uploaded by

Copyright:

Available Formats

Dynamic Simulations in Realistic-Size Networks

Pedro Rafael Barros Martins Arajo

Dissertao para obteno do Grau de Mestre em

Engenharia Electrotcnica e de Computadores

Prof. Doutor Paulo Jos da Costa Branco

Prof. Doutor Pedro Alexandre Flores Correia

Prof. Doutor Jos Manuel Dias Ferreira de Jesus

Palavras Chave: Anlise de sistemas de energia, Estabilidade transitria, Modelos dinmicos,

Context and Motivation ............................................................................................. 2

State of the Art .......................................................................................................... 3

Dissertation Outline ................................................................................................... 6

Power System Simulation Software ................................................................................... 9

Simulation Data Acquisition ............................................................................. 10

Power Flow Computation................................................................................. 10

Modelling of the Network Equations ................................................................ 12

Computation of the Initial Conditions for Dynamic Simulations ........................ 12

Computation of the Algebraic Dynamic State Equations ................................... 13

Dynamic Simulation ......................................................................................... 13

Simulation Software Improvements and Modifications ............................................ 15

Reading of the Dynamic Data Files ................................................................... 15

Presentation of the Results .............................................................................. 17

Generator Reactive Power Limits ..................................................................... 18

Dynamic Models Description ........................................................................................... 21

Round Rotor Synchronous Generator ...................................................................... 22

Excitation Control Systems....................................................................................... 23

Excitation Control System IEEE Type I, IEEET1................................................... 25

Excitation Control System Type DC1A, IEEEX1 .................................................. 26

Simulation Software Algorithm ................................................................................ 10

Governor Control Systems ....................................................................................... 27

Hydro-Turbine Governor, HYGOV ..................................................................... 27

Gas-Turbine Governor, GAST............................................................................ 29

Differential-Algebraic Model ........................................................................................... 33

Modified Euler-Cauchy integration algorithm .......................................................... 34

Models in Differential Form ..................................................................................... 35

GENROU and GENROE model ........................................................................... 35

IEEET1 model ................................................................................................... 36

IEEEX1 model ................................................................................................... 37

HYGOV model .................................................................................................. 38

GAST model ..................................................................................................... 39

GENROU and GENROE models ......................................................................... 41

IEEET1 model ................................................................................................... 42

IEEEX1 model ................................................................................................... 43

HYGOV model .................................................................................................. 44

GAST model ..................................................................................................... 45

Initial Conditions for Dynamic Simulation ................................................................ 46

GENROU and GENROE Initial Conditions........................................................... 46

IEEET1 Initial Conditions ................................................................................... 48

IEEEX1 Initial Conditions................................................................................... 48

HYGOV Initial Conditions .................................................................................. 49

GAST Initial Conditions ..................................................................................... 50

Simulation Results ........................................................................................................... 51

Dynamic Simulation Procedure ................................................................................ 52

Validation of the Dynamic Models ........................................................................... 52

GAST Validation ............................................................................................... 54

IEEEX1 Validation ............................................................................................. 56

IEEET1 Validation ............................................................................................. 59

HYGOV Validation ............................................................................................ 60

Models in Algebraic Form ........................................................................................ 40

57-Bus Case ............................................................................................................. 64

GAST Validation ....................................................................................................... 83