This action might not be possible to undo. Are you sure you want to continue?
Electric Power Systems
Lecture 227052800, ITET ETH
G¨oran Andersson
EEH  Power Systems Laboratory
ETH Z¨ urich
February 2012
ii
Contents
Preface v
1 Introduction 1
1.1 Control Theory Basics  A Review . . . . . . . . . . . . . . . 2
1.1.1 Simple Control Loop . . . . . . . . . . . . . . . . . . . 2
1.1.2 State Space Formulation . . . . . . . . . . . . . . . . . 5
1.2 Control of Electric Power Systems . . . . . . . . . . . . . . . 5
1.2.1 General considerations . . . . . . . . . . . . . . . . . . 5
2 Frequency Dynamics in Electric Power Systems 9
2.1 Dynamic Model of the System Frequency . . . . . . . . . . . 9
2.1.1 Dynamics of the Generators . . . . . . . . . . . . . . . 9
2.1.2 Frequency Dependency of the Loads . . . . . . . . . . 14
2.2 Dynamic Response of Uncontrolled Power System . . . . . . . 17
2.3 The Importance of a Constant System Frequency . . . . . . . 18
2.4 Control Structures for Frequency Control . . . . . . . . . . . 18
3 Primary Frequency Control 21
3.1 Implementation of Primary Control in the Power Plant . . . . 21
3.2 Static Characteristics of Primary Control . . . . . . . . . . . 23
3.2.1 Role of speed droop depending on type of power system 24
3.3 Dynamic Characteristics of Primary Control . . . . . . . . . . 27
3.3.1 Dynamic Model of a OneArea System . . . . . . . . . 27
3.3.2 Dynamic Response of the OneArea System . . . . . . 29
3.3.3 Extension to a TwoArea System . . . . . . . . . . . . 31
3.3.4 Dynamic Response of the TwoArea System . . . . . . 33
3.4 Turbine Modelling and Control . . . . . . . . . . . . . . . . . 36
3.4.1 Turbine Models . . . . . . . . . . . . . . . . . . . . . . 36
3.4.2 Steam Turbine Control Valves . . . . . . . . . . . . . 45
3.4.3 Hydro Turbine Governors . . . . . . . . . . . . . . . . 46
3.5 Dynamic Responses including Turbine Dynamics . . . . . . . 47
iii
iv Contents
4 Load Frequency Control 51
4.1 Static Characteristics of AGC . . . . . . . . . . . . . . . . . . 51
4.2 Dynamic Characteristics of AGC . . . . . . . . . . . . . . . . 54
4.2.1 Onearea system . . . . . . . . . . . . . . . . . . . . . 54
4.2.2 Twoarea system – unequal sizes – disturbance response 56
4.2.3 Twoarea system, unequal sizes – normal control op
eration . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.2.4 Twoarea system – equal sizes, including saturations
– disturbance response . . . . . . . . . . . . . . . . . . 60
5 Synchronous Machine Model 61
5.1 Park’s Transformation . . . . . . . . . . . . . . . . . . . . . . 61
5.2 The Inductance Matrices of the Synchronous Machine . . . . 65
5.3 Voltage Equations for the Synchronous Machine . . . . . . . . 67
5.4 Synchronous, Transient, and Subtransient Inductances . . . . 70
5.5 Time constants . . . . . . . . . . . . . . . . . . . . . . . . . . 74
5.6 Simpliﬁed Models of the Synchronous Machine . . . . . . . . 76
5.6.1 Derivation of the fourthorder model . . . . . . . . . . 76
5.6.2 The HeﬀronPhillips formulation for stability studies . 79
6 Voltage Control in Power Systems 85
6.1 Relation between voltage and reactive power . . . . . . . . . 85
6.2 Voltage Control Mechanisms . . . . . . . . . . . . . . . . . . 87
6.3 Primary Voltage Control . . . . . . . . . . . . . . . . . . . . . 88
6.3.1 Synchronous Machine Excitation System and AVR . . 88
6.3.2 Reactive Shunt Devices . . . . . . . . . . . . . . . . . 93
6.3.3 Transformer Tap Changer Control . . . . . . . . . . . 94
6.3.4 FACTS Controllers . . . . . . . . . . . . . . . . . . . . 95
6.4 Secondary Voltage Control . . . . . . . . . . . . . . . . . . . . 99
7 Stability of Power Systems 101
7.1 Damping in Power Systems . . . . . . . . . . . . . . . . . . . 101
7.1.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . 101
7.1.2 Causes of Damping . . . . . . . . . . . . . . . . . . . . 102
7.1.3 Methods to Increase Damping . . . . . . . . . . . . . . 103
7.2 Load Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . 104
7.2.1 The Importance of the Loads for System Stability . . 104
7.2.2 Load Models . . . . . . . . . . . . . . . . . . . . . . . 104
References 110
A Connection between per unit and SI Units for the Swing
Equation 113
B Inﬂuence of Rotor Oscillations on the Curve Shape 115
Preface
These lectures notes are intended to be used in the lecture Dynamics and
Control of Power Systems (Systemdynamik und Leittechnik der elektrischen
Energieversorgung) (Lecture 227052800, DITET, ETH Z¨ urich) given at
ETH Z¨ urich in the Master Programme of Electrical Engineering and Infor
mation Technology.
The main topic covered is frequency control in power systems. The
needed models are derived and the primary and secondary frequency con
trol are studied. A detailed model of the synchronous machine, based on
Park’s transformation, is also included. The excitation and voltage control
of synchronous machines are brieﬂy described. An overview of load models
is also given.
Z¨ urich, February 2012
G¨oran Andersson
v
vi Preface
1
Introduction
In this chapter a general introduction to power systems control is given.
Some basic results from control theory are reviewed, and an overview of the
use of diﬀerent kinds of power plants in a system is given.
The main topics of these lectures will be
• Power system dynamics
• Power system control
• Security and operational eﬃciency.
In order to study and discuss these issues the following tools are needed
• Control theory (particularly for linear systems)
• Modelling
• Simulation
• Communication technology.
The studied system comprises the subsystems Electricity Generation, Trans
mission, Distribution, and Consumption (Loads), and the associated control
system has a hierarchic structure. This means that the control system con
sists of a number of nested control loops that control or regulate diﬀerent
quantities in the system. In general the control loops on lower system levels,
e.g. locally in a generator, are characterized by smaller time constants than
the control loops active on a higher system level. As an example, the Auto
matic Voltage Regulator (AVR), which regulates the voltage of the generator
terminals to the reference (set) value, responds typically in a time scale of a
second or less, while the Secondary Voltage Control, which determines the
reference values of the voltage controlling devices, among which the genera
tors, operates in a time scale of tens of seconds or minutes. That means that
these two control loops are virtually decoupled. This is also generally true
for other controls in the systems, resulting in a number of decoupled control
loops operating in diﬀerent time scales. A schematic diagram showing the
diﬀerent time scales is shown in Figure 1.1.
1
2 1. Introduction
Protection
VoltageControl
TurbineControl
TieLinePowerand
FrequencyControl
1/10110100 Time(s)
Figure 1.1. Schematic diagram of diﬀerent time scales of power system controls.
The overall control system is very complex, but due to the decoupling
it is in most cases possible to study the diﬀerent control loops individually.
This facilitates the task, and with appropriate simpliﬁcations one can quite
often use classical standard control theory methods to analyse these con
trollers. For a more detailed analysis, one usually has to resort to computer
simulations.
A characteristic of a power system is that the load, i.e. the electric power
consumption, varies signiﬁcantly over the day and over the year. This con
sumption is normally uncontrolled. Furthermore, since substantial parts of
the system are exposed to external disturbances, the possibility that lines
etc. could be disconnected due to faults must be taken into account. The
task of the diﬀerent control systems of the power system is to keep the power
system within acceptable operating limits such that security is maintained
and that the quality of supply, e.g. voltage magnitudes and frequency, is
within speciﬁed limits. In addition, the system should be operated in an
economically eﬃcient way. This has resulted in a hierarchical control sys
tem structure as shown in Figure 1.2.
1.1 Control Theory Basics  A Review
The decoupled control loops described above can be analyzed by standard
methods from the control theory. Just to refresh some of these concepts,
and to explain the notation to be used, a very short review is given here.
1.1.1 Simple Control Loop
The control system in Figure 1.3 is considered. In this ﬁgure the block G(s)
represents the controlled plant and also possible controllers. From this ﬁgure
1.1. Control Theory Basics  A Review 3
System Control Center
State Estimation
Power Flow Control
Economic Dispatch
Security Assessment
Generation
(In Power Stations)
Turbine Control
Voltage Control
Power Transmission
And Distribution
Tap Changer Control
(Direct and Quadature)
Reactive Power
Compensation
HVDC, FACTS
NETWORK
Loads
Normally not
controlled
Figure 1.2. The structure of the hierarchical control systems of a power system.
the following quantities are deﬁned
1
:
• r(t) = Reference (set) value (input)
• e(t) = Control error
• y(t) = Controlled quantity (output)
• v(t) = Disturbance
Normally the controller is designed assuming that the disturbance is equal
to zero, but to verify the robustness of the controller realistic values of v
must be considered.
In principle two diﬀerent problems are solved in control theory:
1. Regulating problem
2. Tracking problem
1
Here the quantities in the time domain are denoted by small letters, while the Laplace
transformed corresponding quantities are denoted by capital letters. In the following this
convention is not always adhered to, but it should be clear from the context if the quantity
is expressed in the time or the s domain.
4 1. Introduction
Σ
e
+
_
G(s)
H(s)
r y
P
v
Figure 1.3. Simple control system with control signals.
In the regulating problem, the reference value r is normally kept constant
and the task is to keep the output close to the reference value even if dis
turbances occur in the system. This is the most common problem in power
systems, where the voltage, frequency, and other quantities should be kept
at the desired values irrespective of load variations, line switchings, etc.
In the tracking problem the task is to control the system so that the
output y follows the time variation of the input r as good as possible. This
is sometimes also called the servo problem.
The transfer function from the input, R, to the output, Y , is given by
(in Laplace transformed quantities)
F(s) =
Y (s)
R(s)
=
C(s)
R(s)
=
G(s)
1 +G(s)H(s)
(1.1)
In many applications one is not primarily interested in the detailed time
response of a quantity after a disturbance, but rather the value directly after
the disturbance or the stationary value when all transients have decayed.
Then the two following properties of the Laplace transform are important:
g(t →0+) = lim
s→∞
sG(s) (1.2)
and
g(t →∞) = lim
s→0
sG(s) (1.3)
where G is the Laplace transform of g. If the input is a step function,
Laplace transform = 1/s, and F(s) is the transfer function, the initial and
stationary response of the output would be
y(t →0+) = lim
s→∞
F(s) (1.4)
and
y(t →∞) = lim
s→0
F(s) (1.5)
1.2. Control of Electric Power Systems 5
1.1.2 State Space Formulation
A linear and timeinvariant controlled system is deﬁned by the equations
_
_
_
˙ x = Ax +Bu
y = Cx +Du
(1.6)
The vector x = (x
1
x
2
. . . x
n
)
T
contains the states of the system, which
uniquely describe the system. The vector u has the inputs as components,
and the vector y contains the outputs as components. The matrix A, of
dimension n × n, is the system matrix of the uncontrolled system. The
matrices B, C, and D depend on the design of the controller and the available
outputs. In most realistic cases D = 0, which means that there is zero
feedthrough, and the system is said to be strictly proper. The matrices A
and B deﬁne which states are controllable, and the matrices A and C deﬁne
which states are observable. A controller using the outputs as feedback
signals can be written as u = −Ky = −KCx, assuming D = 0, where the
matrix K deﬁnes the feedback control, the controlled system becomes
˙ x = (A −BKC)x (1.7)
1.2 Control of Electric Power Systems
1.2.1 General considerations
The overall control task in an electric power system is to maintain the bal
ance between the electric power produced by the generators and the power
consumed by the loads, including the network losses, at all time instants.
If this balance is not kept, this will lead to frequency deviations that if too
large will have serious impacts on the system operation. A complication is
that the electric power consumption varies both in the short and in the long
time scales. In the long time scale, over the year, the peak loads of a day are
in countries with cold and dark winters higher in the winter, so called winter
peak, while countries with very hot summers usually have their peak loads
in summer time, summer peak. Examples of the former are most European
countries, and of the latter Western and Southern USA. The consumption
varies also over the day as shown in Figure 1.4. Also in the short run the
load ﬂuctuates around the slower variations shown in Figure 1.4, so called
spontaneous load variations.
In addition to keeping the above mentioned balance, the delivered elec
tricity must conform to certain quality criteria. This means that the voltage
magnitude, frequency, and wave shape must be controlled within speciﬁed
limits.
If a change in the load occurs, this is in the ﬁrst step compensated by
the kinetic energy stored in the rotating parts, rotor and turbines, of the
6 1. Introduction
holidays
0
20
40
60
80
100
1 4 7 10 13 16 19 22
0
20
40
60
80
100
120
1 4 7 10 13 16 19 22
weekdays
holidays
holidays
weekdays
Time Time
L
o
a
d
(
%
)
L
o
a
d
(
%
)
Figure 1.4. Typical load variations over a day. Left: Commercial
load.; Right: Residential load.
generators resulting in a frequency change. If this frequency change is too
large, the power supplied from the generators must be changed, which is done
through the frequency control of the generators in operation. An unbalance
in the generated and consumed power could also occur as a consequence of
that a generating unit is tripped due to a fault. The task of the frequency
control is to keep the frequency deviations within acceptable limits during
these events.
To cope with the larger variations over the day and over the year gener
ating units must be switched in and oﬀ according to needs. Plans regarding
which units should be on line during a day are done beforehand based on
load forecasts
2
. Such a plan is called unit commitment. When making such
a plan, economic factors are essential, but also the time it takes to bring a
generator online from a state of standstill. For hydro units and gas tur
bines this time is typically of the order of some minutes, while for thermal
power plants, conventional or nuclear, it usually takes several hours to get
the unit operational. This has an impact on the unit commitment and on
the planning of reserves in the system
3
.
Depending on how fast power plants can be dispatched, they are clas
siﬁed as peak load, intermediate load, or base load power plants. This
classiﬁcation is based on the time it takes to activate the plants and on the
2
With the methods available today one can make a load forecast a day ahead which
normally has an error that is less than a few percent.
3
In a system where only one company is responsible for the power generation, the unit
commitment was made in such a way that the generating costs were minimized. If several
power producers are competing on the market, liberalized electricity market, the situation
is more complex. The competing companies are then bidding into diﬀerent markets, pool,
bilateral, etc, and a simple cost minimizing strategy could not be applied. But also in
these cases a unit commitment must be made, but according to other principles.
1.2. Control of Electric Power Systems 7
fuel costs and is usually done as below
4
. The classiﬁcation is not unique and
might vary slightly from system to system.
• Peak load units, operational time 1000–2000 h/a
– Hydro power plants with storage
– Pumped storage hydro power plants
– Gas turbine power plants
• Intermediate load units, operational time 3000–4000 h/a
– Fossil fuel thermal power plants
– Bio mass thermal power plants
• Base load units, operational time 5000–6000 h/a
– Run of river hydro power plants
– Nuclear power plants
In Figure 1.5 the use of diﬀerent power plants is shown in a load duration
curve representing one year’s operation.
The overall goal of the unit commitment and the economic dispatch is
the
• Minimization of costs over the year
• Minimization of fuel costs and start/stop costs
4
The fuel costs should here be interpreted more as the “value” of the fuel. For a hydro
power plant the “fuel” has of course no cost per se. But if the hydro plant has a storage
with limited capacity, it is obvious that the power plant should be used during high load
conditions when generating capacity is scarce. This means that the “water value” is high,
which can be interpreted as a high fuel cost.
8 1. Introduction
System Load
Time
1 year
Run of river hydro power
Nuclear power
Controllable hydro power
Thermal power (fossil fuel)
Hydro power reserves
Gas turbines
Figure 1.5. Duration curve showing the use of diﬀerent kinds of power plants.
2
Frequency Dynamics in Electric Power
Systems
In this chapter, the basic dynamic frequency model for a large power system
is introduced. It is based on the swing equation for the set of synchronous
machines in the system. Certain simpliﬁcations lead to a description of the
dominant frequency dynamics by only one diﬀerential equation which can
be used for the design of controllers. Also of interest is the frequency de
pendency of the load in the system, which has a stabilizing eﬀect on the
frequency. Note that no control equipment is present yet in the models pre
sented in this chapter: the system is shown in ”open loop” in order to un
derstand the principal dynamic behaviour. Control methods are presented in
the subsequent chapters.
2.1 Dynamic Model of the System Frequency
In order to design a frequency control methodology for power systems, the
elementary dynamic characteristics of the system frequency have to be un
derstood. For this purpose, a simpliﬁed model of a power system with sev
eral generators (synchronous machines) will be derived in the sequel. The
nominal frequency is assumed to be 50 Hz as in the ENTSOE Continen
tal Europe system (former UCTE). Generally, deviations from this desired
value arise due to imbalances between the instantaneous generation and con
sumption of electric power, which has an accelerating or decelerating eﬀect
on the synchronous machines.
2.1.1 Dynamics of the Generators
After a disturbance in the system, like a loss of generation, the frequency
in diﬀerent parts of a large power system will vary similar to the exemplary
illustration shown in Figure 2.1. The frequencies of the diﬀerent machines
can be regarded as comparatively small variations over an average frequency
in the system. This average frequency, called the system frequency, is the
frequency that can be deﬁned for the so–called centre of inertia (COI) of the
system.
We want to derive a model that is valid for reasonable frequency devia
tions. For this purpose, the exact version of the swing equation will be used
9
10 2. Frequency Dynamics in Electric Power Systems
50
49.8
49.6
49.4
49
48.8
49.2
0 1 2 3 4 5
f
(
H
z
)
t(s)
Figure 2.1. The frequency in diﬀerent locations in an electric power
system after a disturbance. The thicker solid curve indicates the aver
age system frequency. Other curves depict the frequency of individual
generators.
to describe the dynamic behaviour of generator i:
˙ ω
i
=
ω
0
2H
i
(T
mi
(p.u.) −T
ei
(p.u.)) , (2.1)
with the usual notation. The indices m and e denote mechanical (turbine)
and electrical quantities respectively. ω
i
is the absolute value of the rotor
angular frequency of generator i. The initial condition for eq. (2.1), the
predisturbance frequency, is normally the nominal frequency ω
i
(t
0
) = ω
0
.
Of main interest is usually the angular frequency deviation ∆ω
i
:
∆ω
i
= ω
i
−ω
0
. (2.2)
By deriving eq. (2.2) with respect to the time, one obtains ∆˙ ω
i
= ˙ ω
i
. Note
that for rotor oscillations the frequency of ∆ω
i
is often of interest, while the
amplitude of ∆ω
i
is the main concern in frequency control. Also note that
for the initial condition ∆ω
i
(t
0
) = 0 holds if eq. (2.1) is formulated for ∆ω.
In order to convert the torques in eq. (2.1) to power values, the relation
P(p.u.) = T(p.u.)
ω
ω
0
is used, which yields:
∆˙ ω
i
=
ω
2
0
2H
i
ω
i
(P
mi
(p.u.) −P
ei
(p.u.)) , (2.3)
The power can also be expressed in SI–units (e.g. in MW instead of p.u.) by
multiplication with the power base S
Bi
, which represents the rated power of
2.1. Dynamic Model of the System Frequency 11
G T
e
P
Load
l o a d
P
Generator
Turbine
set
m
P
m
P
Figure 2.2. Simpliﬁed representation of a power system consisting of
a single generator connected to the same bus as the load.
the generator i. Furthermore, eq. (2.3) can be rewritten such that the unit
on both sides is MW:
2H
i
S
Bi
ω
0
∆˙ ω
i
=
ω
0
ω
i
(P
mi
−P
ei
) . (2.4)
Note that this is still the exact version of the swing equation, which is
nonlinear. For further details on diﬀerent formulations of the swing equa
tion, please refer to Appendix A. Now, the goal is to derive the diﬀerential
equation for the entire system containing n generators. In a highly meshed
system, all units can be assumed to be connected to the same bus, represent
ing the centre of inertia of the system. With further simpliﬁcations, they
can even be condensed into one single unit. An illustration of this modelling
is depicted in Figure 2.2. A summation of all the equations (2.4) for the n
generators in the system yields
2
n
i=1
H
i
S
Bi
1
ω
0
∆ ˙ ω
i
=
n
i=1
ω
0
ω
i
(P
mi
−P
ei
) . (2.5)
Because of the strong coupling of the generation units, ω
i
= ω can be as
12 2. Frequency Dynamics in Electric Power Systems
+
–
*/*
2
0
2
B
HS
e 1
s
0
e e e = + A
e
P
m
P
a / a b
b
Figure 2.3. Block diagram of nonlinear frequency dynamics as in eq. (2.11).
sumed for all i. By deﬁning the quantities
ω =
i
H
i
ω
i
H
i
Centre of Inertia frequency (2.6)
S
B
=
i
S
Bi
Total rating, (2.7)
H =
i
H
i
S
Bi
i
S
Bi
Total inertia constant, (2.8)
P
m
=
i
P
mi
Total mechanical power, (2.9)
P
e
=
i
P
ei
Total electrical power, (2.10)
the principal frequency dynamics of the system can be described by the
nonlinear diﬀerential equation
∆˙ ω =
ω
2
0
2HS
B
ω
(P
m
−P
e
) . (2.11)
Eq. (2.11) is illustrated as a block diagram in Figure 2.3. For the frequency
ω in the centre of inertia holds as well
ω = ω
0
+ ∆ω . (2.12)
In order to obtain a linear approximation of eq. (2.11), ω = ω
0
can be
assumed for the righthand side. This is a valid assumption for realistic
frequency deviations in power systems. This yields
∆˙ ω =
ω
0
2HS
B
(P
m
−P
e
) . (2.13)
The dynamics can also be expressed in terms of frequency instead of angular
frequency. Because of ω = 2πf and ˙ ω = 2π
˙
f follows
∆
˙
f =
f
0
2HS
B
(P
m
−P
e
) . (2.14)
2.1. Dynamic Model of the System Frequency 13
A very simple and useful model can be derived if some more assumptions
are made. The overall goal of our analysis is to derive an expression that
gives the variation of ∆ω after a disturbance of the balance between P
m
and
P
e
. Therefore, we deﬁne
P
m
=
i
P
mi
= P
m0
+ ∆P
m
, (2.15)
where P
m0
denotes the mechanical power produced by the generators in
steady state and ∆P
m
denotes a deviation from that value. The total gen
erated power is consumed by the loads and the transmission system losses,
i.e.
P
e
=
i
P
ei
= P
load
+P
loss
, (2.16)
which can, in the same way as in eq. (2.15), be written as
P
e
= P
e0
+ ∆P
load
+ ∆P
loss
(2.17)
with
P
e0
= P
load0
+P
loss0
. (2.18)
If the system is in equilibrium prior to the disturbance,
P
m0
= P
e0
(2.19)
and
P
m0
= P
load0
+P
loss0
(2.20)
are valid. Furthermore, the transmission losses after and before the distur
bance are assumed to be equal, i.e.
∆P
loss
= 0 . (2.21)
If neither the disturbance nor the oscillations in the transmission system are
too large, these approximations are reasonable. Using eqs. (2.15) – (2.21),
eq. (2.13) can now be written as
∆˙ ω =
ω
0
2HS
B
(∆P
m
−∆P
load
) , (2.22)
or equivalently
∆
˙
f =
f
0
2HS
B
(∆P
m
−∆P
load
) . (2.23)
Eq. (2.23) can be represented by the block diagram in Figure 2.4.
14 2. Frequency Dynamics in Electric Power Systems
f '
load
P '
m
P '
System inertia
0
(2 )
B
f
HS s
Figure 2.4. Linearized model of the power system frequency dynamics.
2.1.2 Frequency Dependency of the Loads
Loads are either frequencydependent or frequencyindependent. In real
power systems, a frequency dependency of the aggregated system load is
clearly observable. This has a stabilizing eﬀect on the system frequency f,
as will be shown in the sequel. Apart from a component depending directly
on f, large rotating motor loads cause an additional contribution depending
on
˙
f. This is due to the fact that kinetic energy can be stored in the rotating
masses of the motors.
A load model that captures both eﬀects is given by
P
f
load
−P
f
0
load
= ∆P
f
load
= K
l
∆f +g(∆
˙
f) (2.24)
where
• P
f
0
load
: Load power when f = f
0
,
• K
l
: Frequency dependency,
• g(∆
˙
f): Function that models the loads with rotating masses.
The function g(∆
˙
f) will now be derived. The rotating masses have the
following kinetic energy:
W(f) =
1
2
J(2πf)
2
(2.25)
The change in the kinetic energy, which is equal to the power P
M
consumed
by the motor, is given by
P
M
=
dW
dt
(2.26)
and
∆P
M
=
d∆W
dt
. (2.27)
2.1. Dynamic Model of the System Frequency 15
∆W can be approximated by
W(f
0
+ ∆f) = 2π
2
J(f
0
+ ∆f)
2
=
W
0
+ ∆W = 2π
2
Jf
2
0
+ 2π
2
J2f
0
∆f + 2π
2
J(∆f)
2
= W
0
+
2W
0
f
0
∆f +
W
0
f
2
0
(∆f)
2
⇒∆W ≈
2W
0
f
0
∆f
⇒∆P
M
≈
2W
0
f
0
d∆f
dt
=
2W
0
f
0
∆
˙
f (2.28)
The frequency dependency of the remaining load can also be written as
∂P
load
∂f
∆f = K
l
∆f =
1
D
l
∆f . (2.29)
The values of W
0
and D
l
are obviously highly dependent on the structure
of the load and can be variable over time. Especially W
0
is only a factor
in power systems with large industrial consumers running heavy rotating
machines. The constant D
l
has typical values such that the variation of the
load is equal to 0 . . . 2 % per % of frequency variation.
The block diagram in Figure 2.5 represents the dynamic load model. To
gether with the power system dynamics derived before, we obtain a dynam
ical system with a ”proportional/diﬀerential control” caused by the loads.
However, this eﬀect is too small to be able to keep the frequency within rea
sonable bounds. As we will see in the next section, the absence of any other
control equipment would lead to unacceptable and remaining frequency de
viations even for moderate disturbances.
The power system model derived so far is shown in Figure 2.6.
16 2. Frequency Dynamics in Electric Power Systems
f '
f
load
P '
1
l
D
0
0
2W
s
f
Figure 2.5. Block diagram of the dynamic load model.
load
P '
f '
e
P '
m
P '
f
load
P '
0
0
2W
s
f
Rotating mass loads
Frequencydependent loads
Power generation change
System inertia
System load change
1
l
D
0
(2 )
B
f
HS s
Figure 2.6. Model of power system without control.
2.2. Dynamic Response of Uncontrolled Power System 17
2.2 Dynamic Response of Uncontrolled Power System
Now we will conduct a numerical simulation of the uncontrolled frequency
dynamics after a disturbance. Both loss of generation and loss of load will
be shown, represented by a positive resp. negative step input on the variable
∆P
load
. Table 2.1 displays the parameters used in the simulation. In Figure
2.7, a time plot of the system frequency is shown corresponding to the
diﬀerent disturbances. Note that this result is purely theoretic as such large
frequency deviations could never occur in a real power system because of
various protection mechanisms. However, it illustrates well the possible
frequency rise or decay and the stabilizing selfcontrol eﬀect caused by the
frequency dependency of the load.
Parameter Value Unit
H 5 s
S
B
10 GW
f
0
50 Hz
D
L
1
200
Hz
MW
W
0
100
MW
Hz
Table 2.1. Parameters for time domain simulation of power system.
The plot shows the time evolution of the system frequency for a dis
turbance of (top to bottom) ∆P
load
= −1000 MW, ∆P
load
= −500 MW,
∆P
load
= −100 MW (sudden loss of load) and ∆P
load
= 100 MW, ∆P
load
=
500 MW, ∆P
load
= 1000 MW (sudden increase of load or loss of generation).
0 10 20 30 40 50 60
44
46
48
50
52
54
56
Time [s]
S
y
s
t
e
m
F
r
e
q
u
e
n
c
y
[
H
z
]
−1000 MW
−500 MW
−100 MW
+100 MW
+500 MW
+1000 MW
Figure 2.7. Theoretical frequency responses of uncontrolled power
system (D
L
= 1/200 Hz/MW).
18 2. Frequency Dynamics in Electric Power Systems
2.3 The Importance of a Constant System Frequency
In the most common case ∆P
m
−∆P
load
is negative after a disturbance, like
the tripping of a generator. It is also possible that the frequency rises dur
ing a disturbance, for example when an area that contains much generation
capacity is isolated. Since too large frequency deviations in a system are
not acceptable, automatic frequency control, which has the goal of keeping
the frequency during disturbances at an acceptable level, is used. Further
more, the spontaneous load variations in an electric power system result in
a minute–to–minute variation of up to 2%. This alone requires that some
form of frequency control must be used in most systems.
There are at least two reasons against allowing the frequency to deviate
too much from its nominal value. A non–nominal frequency in the system
results in a lower quality of the delivered electrical energy. Many of the
devices that are connected to the system work best at nominal frequency.
Further, too low frequencies (lower than ≈ 47 − 48 Hz) lead to damaging
vibrations in steam turbines, which in the worst case have to be discon
nected. This constitutes an even worse stress on the system and can lead to
a complete power system collapse. In comparison with thermal units, hydro
power plants are more robust and can normally cope with frequencies down
to 45 Hz.
2.4 Control Structures for Frequency Control
In the following two chapters, the control structures that ensure a constant
system frequency of 50 Hz will be described. The automatic control system
consists of two main parts, the primary and secondary control. Tertiary
control, which is manually activated in order to release the used primary and
secondary control reserves after a disturbance, is not discussed here. This is
due to the fact that the utilization of tertiary control reserves is more similar
to the electricity production according to generations schedules (dispatch),
which are based on economic oﬀline optimizations.
The primary control refers to control actions that are done locally (on
the power plant level) based on the setpoints for frequency and power. The
actual values of these can be measured locally, and deviations from the set
values results in a signal that will inﬂuence the valves, gates, servos, etc. in
a primarycontrolled power plant, such that the desired active power output
is delivered. In primary frequency control, the control task of priority is
to bring the frequency back to (short term) acceptable values. However,
there remains an unavoidable frequency control error because the control
law is purely proportional. The control task is shared by all generators
participating in the primary frequency control irrespective of the location of
the disturbance. Further explanations follow in chapter 3.
2.4. Control Structures for Frequency Control 19
In the secondary frequency control, also called Load Frequency Control,
the power setpoints of the generators are adjusted in order to compensate for
the remaining frequency error after the primary control has acted. Apart
from that, another undesired eﬀect has to be compensated by secondary
control: active power imbalances and primary control actions cause changes
in the load ﬂows on the tielines to other areas, i.e. power exchanges not
according to the scheduled transfers. The secondary control ensures by
a special mechanism that this is remedied after a short period of time.
Note that in this control loop the location of the disturbance is considered
when the control action is determined: only disturbances within its own
control zone (area) are ”seen” by the secondary controller. Note that Load
Frequency Control can also be performed manually as in the Nordel power
system. In the ENTSOE Continental Europe interconnected system, an
automatic scheme is used, which can also be called Automatic Generation
Control. This is further discussed in chapter 4.
The basic control structures described above are depicted in Figure 2.8.
Figure 2.9 shows an illustration of the time spans in which these diﬀerent
control loops are active after a disturbance. Note that primary and sec
ondary control are continuously active also in normal operation of the grid
in order to compensate for small ﬂuctuations. Conversely, the deployment
of tertiary reserves occurs less often.
Underfrequency load shedding is a form of system protection and acts on
timescales well under one second. As the activation of this scheme implies
the loss of load in entire regions, it must only be activated if absolutely
necessary in order to save the system. In the ENTSOE Continental Europe
system, the ﬁrst load shedding stage is activated at a frequency of 49 Hz,
causing the shedding of about 15 % of the overall load. In many systems,
a rotating scheme for how the load should be shed, if that is necessary, is
devised. Such a scheme is often called rotating load shedding.
20 2. Frequency Dynamics in Electric Power Systems
Turbine
Governor
Valves or
Gates
Turbine Generator
Electrical system
 Loads
 Transmission lines
 Other generators
Speed
Automatic
Generation
Control (AGC)
(Secondary
control loop)
Setpoint
Calculation
Power,
Frequency
TieLine Powers
(Internal turbine
control loop)
(Primary control loop)
Setpoint from
power generation
schedule and
tertiary control
identical for synchronous
machines in steady state
System Frequency
Turbine
Controller
Load Shedding
(emergency control)
Figure 2.8. Basic structure of frequency control in electric power systems.
30 s 15 min 60 min
Primary
Control
Secondary
Control
Tertiary
Control
Generation
Rescheduling
Power
Time
Figure 2.9. Temporal structure of control reserve usage after a dis
turbance (terminology according to ENTSOE for region Continental
Europe (former UCTE).
3
Primary Frequency Control
In this chapter, the mechanisms for primary frequency control are illustrated.
As will be shown, they are implemented entirely on the power plant level. In
an interconnected power system, not all generation units need to have pri
mary control equipment. Instead, the total amount of necessary primary
control reserves are determined by statistical considerations and the distri
bution on the power plants can vary. Primary control can also be used in
islanded operation of a single generator. These diﬀerent applications will be
discussed along with their principal static and dynamic characteristics.
3.1 Implementation of Primary Control in the Power
Plant
For a thermal unit, a schematic drawing of the primary control is shown in
Figure 3.1. The turbine governor (depicted here together with the internal
turbine controller) acts on a servomotor in order to adjust the valve through
which the live steam (coming from the boiler with high pressure and high
temperature) ﬂows to the turbines. In the high pressure turbine, part of the
energy of the steam is converted into mechanical energy. Often the steam is
then reheated before it is injected into a medium pressure or low pressure
turbine, where more energy is extracted from the steam. In practice, these
turbinegenerator systems can be very large. In a big thermal unit of rating
1000 MW, the total length of the turbinegenerator shaft may exceed 50 m.
The control law, which is shown as a block diagram in Figure 3.2, is a
proportional feedback control. It establishes an aﬃne relation between the
measured frequency and the power generation of the plant in steady state.
Note that the turbine dynamics plays in important role in the overall
dynamical response of the system. However, we will neglect this for the
sake of simplicity in the ﬁrst three sections of this chapter. In section 3.4.1,
diﬀerent steam and hydro turbines and their modelling are discussed and
later their eﬀect on the dynamical response will be shown.
Primary control is implemented on a purely local level; there is no co
ordination between the diﬀerent units. This is also why the control law
cannot have an integral component: integrators of diﬀerent power plants
could start ”competing” each other for power production shares, which can
lead to an unpredictable and unreasonable distribution of power generation
on the available plants.
21
22 3. Primary Frequency Control
Measured values
LP
Electricity grid
Electrical
power
Mech.
power
HP
Actuation
Steam
Steam
Control
signals
Reference values
G
Boiler
Valve
Servomotor
Controller/
Governor
Reheater
0 0
, f P
, f P
Figure 3.1. Schematic drawing of the primary control installed in
a thermal unit. HP = High Pressure Turbine. LP = Low Pressure
turbine.
G T
Proportional
control law
1/ K S
0
set
f f f '
set
m
P '
0
set
m
P
f
, set tot
m
P
Internal
Turbine
Controller
, set tot
m m
P P H
e
P
Turbine shaft
Figure 3.2. Block diagram describing the primary control law. The
measured power value corresponds to P
m
in our notation, while in
practice the measurement can be done on the electrical side.
3.2. Static Characteristics of Primary Control 23
3.2 Static Characteristics of Primary Control
First, it is of particular interest to study the properties of the primary fre
quency control in steady state. Referring to the block diagram in Figure 3.2,
the equation describing the primary control is given by
(f
0
−f) ·
1
S
+ (P
set
m0
−P
set,tot
m
) = 0 , (3.1)
which can be written as
S = −
f
0
−f
P
set
m0
−P
set,tot
m
= −
f −f
0
P
set,tot
m
−P
set
m0
Hz/MW > 0 (3.2)
or in per unit as
S = −
f −f
0
f
0
P
set,tot
m
−P
set
m0
P
set
m0
. (3.3)
Under the assumption that the turbine power controller has an integrating
characteristic (ε → 0 when t → ∞ in Figure 3.2), it follows that in steady
state holds P
m
= P
set,tot
m
.
The speed droop characteristic, Figure 3.3, yields all possible steady
state operating points (P
set,tot
m
, f) of the turbine. The position and slope
of the straight line can be ﬁxed by the parameters P
set
m0
, f
0
and S. We
have chosen to label the horizontal axis with the power P
set,tot
m
which for
small deviations of the frequency around the nominal value is identical to
the torque T. In the literature the speed droop characteristics is sometimes
also described by ω instead of by f.
0
( . .) f pu
,
( . .)
set tot
m
P pu
0
( . .)
set
m
P pu
( . .) f pu
( . .) S pu
Figure 3.3. Static characteristic of primary control.
24 3. Primary Frequency Control
3.2.1 Role of speed droop depending on type of power system
We will now study how the frequency control of a generator will act in
three diﬀerent situations. First, when the generator is part of a large inter
connected system, and second when the generator is in islanded operation
feeding a load. The third system to be studied is a two machine system.
Generator in Large System If a generator is embedded in a large inter
connected system, it can be modelled with a very good approximation as
connected to an inﬁnite bus. This is shown in Figure 3.4.
In steady state the frequency is given by the grid frequency f
G
(repre
sented by the inﬁnite bus). From the speed droop characteristics, Figure 3.5,
the power produced by the generator can then be determined. The turbine
governor controls thus only the power, not the frequency, see Figure 3.5.
0 0
,
set
m
f P
,
G
f P
L
X
e
P
m
P
T
Infinite
Bus
G
f f =
G
U U =
G
R
Figure 3.4. Generator operating in a large interconnected system.
f
0
set
m
P
0
f
S
,
( )
set tot
m G
P f
0 0
und are set in the generator
set
m
f P
( )
G
P g f
, set tot
m
P
G
f
Figure 3.5. Speed droop characteristics for the case when the gener
ator is connected to an inﬁnite bus (large system).
3.2. Static Characteristics of Primary Control 25
Islanded Operation As depicted in Figure 3.6, the generator feeds a load in
islanded operation, which here is assumed to be a purely resistive load. By
a voltage controller the voltage U is kept constant and thus also P
e
. In this
case the primary control loop will control the frequency, not the power. The
resulting frequency can be determined from the speed droop characteristics,
Figure 3.7.
0 0
,
set
m
f P
, f P
e
P
m
P
T
U
R
2
.
e
U
P const
R
= =
G
R
Voltage Control
Figure 3.6. Generator in islanded operation.
f
0
set
m
P
0
f
S
, set tot
m e
P P
0 0
und are set in the generator
set
m
f P
( ) f h P
, set tot
m
P
f
Figure 3.7. Speed droop characteristics for the case when the gener
ator is in islanded operation.
Two Generator System The two generator system, Figure 3.8, provides a
simple model that is often used to study the interaction between two areas
in a large system. In this model the two generators could represent two
subsystems, and the speed droop is then the sum of all the individual speed
droops of the generators in the two subsystems, Figure 3.9. With the help of
the speed droop characteristics of the two systems, we will determine how a
change in load will be compensated by the two systems. Thus, if we have a
change ∆P
load
of the overall load, what will the changes in P
m,1
, P
m,2
, and
f be?
26 3. Primary Frequency Control
This will be solved in the following way:
• The quantities (P
set
m0,1
, f
0,1
, S
1
) and (P
set
m0,2
, f
0,2
, S
2
) describe the speed
droop characteristics of the two systems g
1
and g
2
.
• From these the sum g
3
= g
1
+g
2
is formed.
• From the given P
load
we can determine P
set
m,1
, P
set
m,2
and f
N
from g
3
.
• In a similar way: From P
load
+∆P
load
can P
set
m0,1
+∆P
set
m,1
, P
set
m0,2
+∆P
set
m,2
and f
N
+ ∆f be determined, and thus ∆P
set
m,1
, ∆P
set
m,2
und ∆f.
All these steps are shown in Figure 3.9.
G
2
G
1
,1 ,2
at
load m m N
P P P f
0,1 0,1 1
, ,
set
m
P f S 0,2 0,2 2
, ,
set
m
P f S
,1 m
P
,2 m
P
,1 m
P '
,2 m
P '
Figure 3.8. Two generator system.
1 1
, g S
2 2
, g S
3 3
, g S
0,2
f
0,1
f
N
f f + A
0,2
set
m
P
0,1
set
m
P
,1
set
m
P
,2
set
m
P load
P
load load
P P + A
,1 ,1
set set
m m
P P + A ,2 ,2
set set
m m
P P + A
N
f
Figure 3.9. Speed droop characteristics for a two machine system.
3.3. Dynamic Characteristics of Primary Control 27
3.3 Dynamic Characteristics of Primary Control
3.3.1 Dynamic Model of a OneArea System
In this section, we are going to extend the dynamic frequency model intro
duced in Chapter 2 by the primary control loop in the power plants. For an
individual generator, the block diagram has been introduced in the previous
section.
Following the nomenclature introduced in Figure 3.2, we start with
P
set,tot
m
= P
set
m0
+ ∆P
set
m
= P
set
m0
−
1
S
∆f , (3.4)
where P
set,tot
m
describes the power setpoint of the turbine including the sched
uled value P
set
m0
and the component imposed by primary control. For the
linearized consideration of the power system in ∆ quantities, we set
P
set
m0
= ∆P
set
m0
(3.5)
as the steadystate component cancels out against the other steadystate
quantities. The remaining question is how a change in ∆P
set,tot
m
translates
into an actual mechanical power output change ∆P
m
. Thus, we regard now
the internal turbine control loop as depicted in Figure 3.10.
Turbine
( )
t
G s
Turbine
Controller
t
K
s
set
m
P '
m
P '
0
set
m
P '
Figure 3.10. Block diagram of the turbine and turbine control dynamics.
From this ﬁgure it follows that
∆P
m
(s) =
G
t
(s)
G
t
(s) +
1
K
t
s
_
∆P
set
m0
(s) + ∆P
set
m
(s)
¸
. (3.6)
If the dynamics of the turbine is neglected (G
t
(s) = 1), one obtains
∆P
m
(s) =
1
1 +T
t
s
_
∆P
set
m0
(s) + ∆P
set
m
(s)
¸
. (3.7)
Note that the time constant T
t
= 1/K
t
is fairly small compared with the
frequency dynamics of the system. Regarding now only the frequency control
component (∆P
set
m0
= 0), we obtain in steady state as before
∆P
m
= −
1
S
∆f . (3.8)
28 3. Primary Frequency Control
For the case of n controllers for n generators we obtain analogously for
controller i
∆P
mi
= −
1
S
i
∆f i = 1, . . . , n (3.9)
i
∆P
mi
= −
i
1
S
i
∆f (3.10)
where
∆P
m
=
i
∆P
mi
(3.11)
is the total change in turbine power. By deﬁning
1
S
=
i
1
S
i
(3.12)
we thus have
∆P
m
= −
1
S
∆f . (3.13)
This can be inserted into the dynamic system frequency model derived in
Chapter 2 as depicted in Figure 3.11.
load
P '
f '
e
P '
f
load
P '
0
0
2W
s
f
Rotating mass loads
Frequencydependent loads
System inertia
System load change
1
l
D
0
(2 )
B
f
HS s
m
P '
1
1
t
T s
set
m
P '
0
set
m
P '
1
S
Turbine
dynamics/
control
Primary
control
Figure 3.11. Dynamic frequency model of the power system with
primarycontrolled power plants.
3.3. Dynamic Characteristics of Primary Control 29
3.3.2 Dynamic Response of the OneArea System
Now we are going to study the eﬀect of a disturbance in the system derived
above. Both loss of generation and loss of load can be simulated by imposing
a positive or negative step input on the variable ∆P
load
. A change of the set
value of the system frequency f
0
is not considered as this is not meaningful
in real power systems.
From the block diagram in Figure 3.11 it is straightforward to derive the
transfer function between ∆P
load
and ∆f (∆P
set
m0
= 0):
∆f(s) = −
1 +sT
t
1
S
+
1
D
l
(1 +sT
t
) + (
2W
0
f
0
+
2HS
B
f
0
)s(1 +sT
t
)
∆P
load
(s) (3.14)
The step response for
∆P
load
(s) =
∆P
load
s
(3.15)
is given in Figure 3.12. The frequency deviation in steady state is
∆f
∞
= lim
s→0
(s · ∆f(s)) =
−∆P
load
1
S
+
1
D
l
=
−∆P
load
1
D
R
= −∆P
load
· D
R
(3.16)
with
1
D
R
=
1
S
+
1
D
l
(3.17)
In order to calculate an equivalent time constant T
eq
, T
t
is put to 0. This
can be done since for realistic systems the turbine controller time constant
T
t
is much smaller than the time constant of the frequency dynamics T
M
:
T
t
≪T
M
=
f
0
S
B
(
2W
0
f
0
+
2HS
B
f
0
) . (3.18)
This means that the transfer function in eq. (3.14) can be approximated by
a ﬁrst order function
∆f(s) =
−∆P
load
(s)
1
D
R
+T
M
S
B
f
0
s
=
−1
1 +T
M
D
R
S
B
f
0
s
D
R
∆P
load
s
(3.19)
or
∆f(s) =
1
1 +T
M
D
R
S
B
f
0
s
∆f
∞
s
(3.20)
with
T
eq
= T
M
D
R
S
B
f
0
(3.21)
as the equivalent time constant.
30 3. Primary Frequency Control
0 1 2 3 4 5 6 7 8 9 10
49.75
49.8
49.85
49.9
49.95
50
S
y
s
t
e
m
F
r
e
q
u
e
n
c
y
[
H
z
]
0 1 2 3 4 5 6 7 8 9 10
0
100
200
300
400
Time [s]
P
o
w
e
r
[
M
W
]
∆ P
load
∆ P
m
∆ P
load
f
Figure 3.12. Behaviour of the onearea system (no turbine dynamics,
parameterized as described in the example below) after a step increase
in load. The upper plot shows the system frequency f. The lower plot
shows the step function in ∆P
load
, the increase in turbine power ∆P
m
,
and the frequencydependent load variation ∆P
f
load
.
Example
• S
B
= 4000 MW
• f
0
= 50 Hz
• S = 4% = 0.04
f
0
S
B
=
0.04·50
4000
Hz/MW =
1
2000
Hz/MW
• D
l
= 1 %/1 % ⇒D
l
=
f
0
S
B
=
50
4000
Hz/MW =
1
80
Hz/MW
• ∆P
load
= 400 MW
• T
M
= 10 s
Then follows
∆f
∞
= −D
R
· ∆P
load
= −
1
2000
MW
Hz
+ 80
MW
Hz
· 400MW = −0.19 Hz (3.22)
and
T
eq
= 10 s ·
1
2000
MW
Hz
+ 80
MW
Hz
4000 MW
50 Hz
= 0.39 s . (3.23)
3.3. Dynamic Characteristics of Primary Control 31
3.3.3 Extension to a TwoArea System
Up to now we have mostly studied the behaviour of a power system consist
ing of a single area. If the power system is highly meshed in this case, it can
be represented by a single bus where all units are connected. In practice,
however, a large interconnected power system is always divided into various
”control zones” or ”areas”, corresponding e.g. to countries. Understanding
the interactions between these areas is therefore highly important for the
ﬂawless operation of the entire system. For simpliﬁed simulation studies, a
system with two areas can be represented by two single bus systems with a
tieline in between them. This is depicted in Figure 3.13.
G
1
0,1
set
m
P
T
1
1
f
,1 e
P
Load 1
,1 l o a d
P
1 2 T
P
Tieline
power flow
G
2
T
2
2
f
,1 m
P
, 2 e
P
, 2 m
P
0,2
set
m
P
Load 2
, 2 l o a d
P
Area 1 Area 2
R
1
R
2
Figure 3.13. Simpliﬁed representation of a power system with two areas.
In order to adapt our dynamic frequency model accordingly, the power
exchange P
T12
over the tie line between the areas 1 and 2 has to be modelled.
This is given by
P
T12
=
U
1
U
2
X
sin(ϕ
1
−ϕ
2
) (3.24)
where X is the (equivalent) reactance of the tie line. For small deviations
(U
1
and U
2
are constant) one gets
∆P
T12
=
∂P
T12
∂ϕ
1
∆ϕ
1
+
∂P
T12
∂ϕ
2
∆ϕ
2
=
U
1
U
2
X
cos(ϕ
0,1
−ϕ
0,2
)(∆ϕ
1
−∆ϕ
2
)
(3.25)
or
∆P
T12
=
ˆ
P
T
(∆ϕ
1
−∆ϕ
2
) (3.26)
with
ˆ
P
T
=
U
1
U
2
X
cos(ϕ
0,1
−ϕ
0,2
) . (3.27)
By using this model, the block diagram of the power system can be extended
as shown in Figure 3.14.
32 3. Primary Frequency Control
,1 load
P A
1
f A
÷
,1 e
P A
+
+
+
,1
f
load
P A
+
+
+
+
,2 load
P A
2
f A
÷
,2 e
P A
+
+
+
,2
f
load
P A
+
+
+
12 T
P A
+
÷
12 T
P A
12 T
P A
+
+
÷
0,1
0
2W
s
f
Rotating mass loads
Frequencydependent loads
System inertia
,1
1
l
D
0
1 ,1
(2 )
B
f
H S s
,1 m
P A
,1
1
1
t
T s +
,1
set
m
P A
0,1
set
m
P A
1
1
S
÷
Turbine
dynamics/
control
Primary
control
0,2
0
2W
s
f
Rotating mass loads
Frequencydependent loads
System inertia
,2
1
l
D
0
2 ,2
(2 )
B
f
H S s
,2 m
P A
1
1
t
T s +
,2
set
m
P A
0,2
set
m
P A
2
1
S
÷
Turbine
dynamics/
control
Primary
control
ˆ
2
T
P
s
t
ˆ
2
T
P
s
t
Figure 3.14. Twoarea dynamic model including tieline ﬂows.
3.3. Dynamic Characteristics of Primary Control 33
3.3.4 Dynamic Response of the TwoArea System
For the case regarded here, it is assumed that one of the areas is much smaller
than the other. The bigger one of the two areas can then be regarded as an
inﬁnite bus in our analysis. We will now study the behaviour after a load
change in the smaller area, Area 1. The system to be studied is depicted in
Figure 3.15.
12 T
P
X
1 1
, U ¢
2 2
, U ¢
1
f
2 0
f f =
Area 1
1 ,1 0,1
,1 ,1 0
, ,
, ,
l
M B
S D W
T S f
Area 2
2 ,2 0,2
,2 ,2 0
, ,
, ,
l
M B
S D W
T S f
Figure 3.15. Model of a two area system. Area 1 is much smaller than Area 2.
As Area 2 is very big (inﬁnite bus) it follows that
T
M,2
S
B2
f
0
≫
T
M,1
S
B1
f
0
⇒f
2
= constant ⇒∆ϕ
2
= 0 (3.28)
and consequently
∆P
T12
= −∆P
T21
=
ˆ
P
T
∆ϕ
1
= 2π
ˆ
P
T
_
∆f
1
dt (3.29)
Without any scheduled generator setpoint changes, i.e. ∆P
set
m0,i
= 0, the
following transfer functions apply for a change in the Area 1 system load
∆P
load,1
(s):
∆f
1
(s) =
−s
2π
ˆ
P
T
+ (
1
D
l,1
+
1
S
B,1
(1 +sT
t
)
)s +
T
M,1
S
B,1
f
0
s
2
∆P
load,1
(s)
(3.30)
∆P
T12
(s) =
2π
ˆ
P
T
s
∆f
1
(s) (3.31)
∆P
T12
(s) =
−2π
ˆ
P
T
2π
ˆ
P
T
+ (
1
D
l,1
+
1
S
B,1
(1 +sT
t
)
)s +
T
M,1
S
B,1
f
0
s
2
∆P
load,1
(s)
(3.32)
34 3. Primary Frequency Control
The response for
∆P
load,1
(s) =
∆P
load,1
s
(3.33)
is shown if Figure 3.16. The steady state frequency deviation is
∆f
1,∞
= lim
s→0
(s · ∆f
1
(s)) = 0 (3.34)
and the steady state deviation of the tie line power is
∆P
T12,∞
= lim
s→0
(s · ∆P
T12
(s)) = −∆P
load,1
(3.35)
The inﬁnite bus brings the frequency deviation ∆f
1
back to zero. This
is achieved by increasing the tieline power so the load increase is fully
compensated. While this is beneﬁcial for the system frequency in Area 1,
a new, unscheduled and persisting energy exchange has arisen between the
two areas.
0 1 2 3 4 5 6 7 8 9 10
49.85
49.9
49.95
50
Time [s]
F
r
e
q
u
e
n
c
y
[
H
z
]
f
1
f
2
0 1 2 3 4 5 6 7 8 9 10
0
200
400
600
800
1000
Time [s]
P
o
w
e
r
∆ P
load,1
∆ P
m,1
∆ P
T12
Figure 3.16. Step response for the system in Figure 3.14 resp. Figure
3.15. Only primary control is used and the system is parameterized
according to Table 3.1. The upper diagram shows the frequencies f
1
in Area 1 and f
2
in Area 2. The lower diagram shows the step in the
system load ∆P
load,1
, the turbine power ∆P
m,1
and the tieline power
∆P
T21
= −∆P
T12
.
3.3. Dynamic Characteristics of Primary Control 35
Parameter Value Unit
H
1
5 s
H
2
5 s
S
B,1
10 GW
S
B,2
10 TW
f
0
50 Hz
D
l,1
1
200
Hz
MW
D
l,2
1
200
Hz
TW
W
0,1
0
MW
Hz
W
0,2
0
MW
Hz
S
1
1
5000
Hz
MW
S
2
1
5000
Hz
MW
ˆ
P
T
533.33 MW
∆P
load,1
1000 MW
Table 3.1. Parameters for time domain simulation of the twoarea
power system corresponding to Figure 3.16.
36 3. Primary Frequency Control
3.4 Turbine Modelling and Control
This section gives an overview of the modelling of both steam and hydro
turbines, control valves and governors. Their characteristics and behaviour
are also brieﬂy discussed. The aim here is to give an understanding of the
basic physical mechanisms behind these models that are very commonly
used in simulation packages for the study of power systems dynamics. In
the dynamic frequency model of the power system derived so far (as depicted
in Figure 3.11), the linearized dynamic models of the turbines will enter in
the block ”Turbine dynamics/control”.
3.4.1 Turbine Models
Steam Turbines
Figures 3.17, 3.18, and 3.19 show the most common steam turbines and
their models. It is outside the scope of these lecture notes to give a de
tailed derivation and motivation of these models, only a brief qualitative
discussion will be provided. In a steam turbine the stored energy of high
temperature and high pressure steam is converted into mechanical (rotat
ing) energy, which then is converted into electrical energy in the generator.
The original source of heat can be a furnace ﬁred by fossil fuel (coal, gas, or
oil) or biomass, or it can be a nuclear reactor.
The turbine can be either tandem compound or cross compound. In
a tandem compound unit all sections are on the same shaft with a single
generator, while a cross compound unit consists of two shafts each connected
to a generator. The cross compound unit is operated as one unit with one
set of controls. Most modern units are of tandem compound type, even if
the crossover compound units are more eﬃcient and have higher capacity.
However, the costs are higher and could seldom be motivated.
The power output from the turbine is controlled through the position
of the control valves, which control the ﬂow of steam to the turbines. The
valve position is inﬂuenced by the output signal of the turbine controller.
Following the nomenclature introduced in Figure 3.10, this signal is deﬁned
as
∆P
ctrl
m
=
1
T
t
s
(∆P
set
m0
+ ∆P
set
m
−∆P
m
) . (3.36)
The delay between the diﬀerent parts of the steam path is usually modelled
by a ﬁrst order ﬁlter as seen in Figures 3.17, 3.18, and 3.19. Certain fractions
of the total power are extracted in the diﬀerent turbines, and this is modelled
by the factors F
V HP
, F
HP
, F
IP
, F
LP
in the models. Typical values of the
time constant of the delay between the control valves and the highpressure
turbine, T
CH
, is 0.1 − 0.4 s. If a reheater is installed, the time delay is
larger, typically T
RH
= 4 − 11 s. The time constant of the delay between
the intermediate and low pressure turbines, T
CO
, is in the order of 0.3−0.6 s.
3.4. Turbine Modelling and Control 37
Shaft
To Condenser
ctrl
m
P A
m
P A
Shaft
To Condenser
E E
Valve
position
ctrl
m
P A
m
P A
Control
Valves,
Steam
Chest
HP
1
1
CH
sT +
Nonreheat Steam Turbines
Linear Model
Control
Valves,
Steam
Chest
HP
Reheater
Crossover
IP
LP LP
Tandem Compound, Single Reheat Steam Turbines
1
1
CH
sT +
1
1
RH
sT +
1
1
CO
sT +
HP
F
IP
F LP
F
Linear Model
Valve
Position
Figure 3.17. Steam turbine conﬁgurations and approximate linear
models. Nonreheat and tandem compound, single reheat conﬁgura
tions.
38 3. Primary Frequency Control
Shaft
To Condenser
6 6
Valve
position
6
Control
Valves,
Steam
Chest
VHP
Reheater
Crossover
HP
LP LP
Tandem Compound, Double Reheat Steam Turbine
1
1
CH
sT 1
1
1
RH
sT
2
1
1
RH
sT
HP
F
IP
F
Linear Model
Reheater
IP
1
1
CO
sT
HP
F
ctrl
m
P '
m
P '
VHP
F
Figure 3.18. Steam turbine conﬁgurations and approximate linear
models. Tandem compound, double reheat conﬁguration.
3.4. Turbine Modelling and Control 39
HP, LP Shaft
Valve
position
IP, LP Shaft
6
6
Control
Valves,
Steam
Chest
HP
Reheater Crossover
IP
LP LP
Cross Compound, Single Reheat Steam Turbine
LP LP
1
1
CH
sT
1
1
RH
sT
1
1
CO
sT
HP
F
2
LP
F
Linear Model
2
LP
F
IP
F
ctrl
m
P '
1 m
P '
2 m
P '
Figure 3.19. Steam turbine conﬁgurations and approximate linear
models. Cross compound, single reheat conﬁguration.
40 3. Primary Frequency Control
Step Response To illustrate the dynamics of a steam turbine, the conﬁgu
ration with tandem compound, single reheat, Figure 3.17, with the following
data will be studied:
T
CH
= 0.1 s, T
RH
= 10 s, T
CO
= 0.3 s
F
HP
= 0.3, F
IP
= 0.4, F
LP
= 0.3
As T
CH
≪ T
RH
und T
CO
≪ T
RH
, we can assume T
CH
= T
CO
= 0 for an
approximate analysis. Then a simpliﬁed block diagram according to Fig
ure 3.20 can be used. For this system the step response is easy to calculate.
It is depicted in Figure 3.21.
6
1
1 10 s
0.3 0.7
ctrl
m
P '
m
P '
Figure 3.20. Simpliﬁed model of tandem compound, single reheat
system in Figure 3.17.
0 5 10 15 20 25 30 35 40 45 50
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Time (s)
P
o
w
e
r
f
r
o
m
t
u
r
b
i
n
e
FIP + FLP = 0.7
FHP = 0.3
TRH = 10 s
Figure 3.21. Step response of system in Figure 3.20.
3.4. Turbine Modelling and Control 41
Area A =
Velocity v =
P
1
Length of
Penstock L =
h Head =
P
2
Effective Area a =
Output Velocity v
out
=
Figure 3.22. Schematic drawing of hydro turbine with water paths.
Hydro Turbines
Compared with steam turbines, hydro turbines are easier and cheaper to
control. Thus, frequency control is primarily done in the hydro power plants
if available. If the amount of hydro–generated power in a system is not
suﬃcient, the steam turbines have to be included in the frequency control.
The power produced by a generator is determined by the turbine gover
nor and the dynamic properties of the turbine. Thus, to be able to determine
the frequency’s dynamic behaviour, models for the turbine as well as for the
turbine control are necessary.
Figure 3.22 depicts a hydro turbine with penstock and hydro reservoir
and deﬁnes the notation that will be used from now on. Bernoulli’s equation
for a trajectory between the points P
1
and P
2
can be written as
_
P
2
P
1
∂v
∂t
· dr +
1
2
(v
2
2
−v
2
1
) + Ω
2
−Ω
1
+
_
P
2
P
1
1
ρ
dp = 0 . (3.37)
The following assumptions are usually made:
• v
1
= 0, since the reservoir is large and the water level does not change
during the time scale that is of interest here.
42 3. Primary Frequency Control
• The water velocity is non–zero only in the penstock.
• The water is incompressible, i.e. ρ does not change with water pressure.
• The water pressure is the same at P
1
and P
2
, i.e. p
1
= p
2
.
Further,
Ω
2
−Ω
1
= −gh . (3.38)
The above assumptions together with eq. (3.38) make it possible to write
(3.37), with v
out
= v
2
and the length of the penstock L, as
L
dv
dt
+
1
2
v
2
out
−gh = 0 . (3.39)
The velocity of the water in the penstock is v. The eﬀective opening of the
penstock, determined by the opening of the turbine’s control valve (guide
vanes), is denoted a. If the penstock’s area is A,
v
out
=
A
a
v (3.40)
is valid and eq. (3.39) can be written as
dv
dt
=
1
L
gh −
1
2L
_
A
a
v
_
2
. (3.41)
The maximum available power at the turbine is
P =
1
2
ρav
3
out
=
1
2
ρ
A
3
v
3
a
2
. (3.42)
To get the system into standard form,
_
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
_
x = v ,
u =
a
A
,
y = P ,
(3.43)
are introduced. (Here, we have used the standard notation, i.e. x for state, u
for control signal, and y for output signal.) The system now can be written
as
_
¸
¸
¸
_
¸
¸
¸
_
˙ x =
gh
L
−x
2
1
2Lu
2
,
y = ρA
x
3
2u
2
.
(3.44)
3.4. Turbine Modelling and Control 43
*/*
Σ
u
x
x
x u ⁄
x
2
u
2
⁄
t d
dx
1
2L

1
s

ρA
2

gh
L

y P ρA
x
3
2u
2
 = =
y

+
Figure 3.23. Block diagram showing model of hydro turbine.
The system corresponding to eq. (3.44) can be described with the block
diagram in Figure 3.23.
Eq. (3.44) is nonlinear and a detailed analysis is beyond the scope of
these lecture notes. To get an idea of the system’s properties, the equations
are linearised, and small variations around an operating point are studied.
In steady state, ˙ x = 0, and the state is determined by x
0
, u
0
, and y
0
, which
fulﬁls
_
¸
¸
_
¸
¸
_
x
0
= u
0
_
2gh ,
y
0
=
ρAx
3
0
2u
2
0
.
(3.45)
Small deviations ∆x, ∆u, and ∆y around the operating point satisfy
_
¸
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
¸
_
∆˙ x = −2x
0
1
2Lu
2
0
∆x +
2x
2
0
2Lu
3
0
∆u ,
∆y = 3ρ
Ax
2
0
2u
2
0
∆x −2ρ
Ax
3
0
2u
3
0
∆u ,
(3.46)
which, using eqs. (3.45), can be written as
_
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
_
∆˙ x = −
√
2gh
u
0
L
∆x +
2gh
u
0
L
∆u ,
∆y =
3y
0
u
0
√
2gh
∆x −
2y
0
u
0
∆u .
(3.47)
44 3. Primary Frequency Control
∆u
0 5 10 15 20
∆u
1
t s ( )
∆y
y
0
u
0
∆u
1
2
y
0
u
0
  ∆u
1
t s ( )
T
w
5 s =
Figure 3.24. The variation of the produced power, ∆y, after a step
change in the control valve.
The quantity L/
√
2gh has dimension of time, and from the above equa
tions it is apparent that this is the time it takes the water to ﬂow through
the penstock if a = A. That time is denoted T:
T = L/
_
2gh . (3.48)
If eqs. (3.47) are Laplace–transformed, ∆x can be solved from the ﬁrst
of the equations, leading to
∆x =
L/T
1 +su
0
T
∆u , (3.49)
which, when inserted in the lower of eqs. (3.47) gives
∆y =
y
0
u
0
·
1 −2u
0
Ts
1 +u
0
Ts
∆u . (3.50)
The quantity u
0
T = a
0
T/A also has dimension of time and is denoted T
w
.
The time constant T
w
is the time it takes for the water to ﬂow through the
penstock when a = a
0
or u = u
0
, i.e. for the operation point where the
linearization is done. Eq. (3.50) can thus be written as
∆y =
y
0
u
0
·
1 −2T
w
s
1 +T
w
s
∆u . (3.51)
3.4. Turbine Modelling and Control 45
It is evident that the transfer function in eq. (3.51) is of non–minimum
phase, i.e. not all poles and zeros are in the left half plane. In this case, one
zero is in the right half plane. That is evident from the step response to
eq. (3.51), depicted in Figure 3.24.
The system has the peculiar property to give a lower power just after
the opening of the control valve is increased before the desired increased
power generation is reached. The physical explanation is the lower pressure
appearing after the control valve is opened, so that the water in the penstock
can be accelerated. When the water has been accelerated, the generated
power is increased as a consequence of the increased ﬂow. That property of
water turbines places certain demands on the design of the control system
for the turbines.
3.4.2 Steam Turbine Control Valves
As stated earlier, the control input of the steam turbine acts on a valve which
inﬂuences the inﬂow of live steam into the turbine. As the valve cannot be
moved at inﬁnite speed, a further dynamics is introduced into the system.
This is usually modelled by a ﬁrstorder element. Furthermore, constraints
exist on the ramp rate (speed of the valve motion) and the absolute valve
position. Note that, if the block diagram is expressed in ∆ quantities, the
latter has to be formulated in ∆ quantities as well, taking into account the
steadystate valve position. Figure 3.25 depicts the corresponding block
diagram.
Note that this block must be inserted between the turbine controller
output ∆P
ctrl
m
and the turbine input when the valve dynamics shall be con
sidered. The output of the valve can be named then ∆P
ctrl
m
and its input
could be renamed to e.g. ∆P
ctrl
∗
m
.
1
V
T
1
s
* ctrl
m
P '
ctrl
m
P '
up
P '
down
P '
,0 min min min
P P P ' '
,0 max max max
P P P '
Figure 3.25. Model of a control valve.
46 3. Primary Frequency Control
3.4.3 Hydro Turbine Governors
The task of a turbine governor is to control the turbine power output such
that it is equal to a set value which consists of a constant and a frequency
dependent part, the latter of which is of interest here. For steam turbines,
this has been discussed in Section 3.1. Because of the particularities of hydro
turbines, which have been shown in Section 3.4.1, another type of governor
is needed.
A model of a hydro turbine governor is given in Figure 3.26. The control
servo is here represented simply by a time constant T
p
. The main servo,
i.e. the guide vane, is represented by an integrator with the time constant
T
G
. Typical values for these parameters are given in Table 3.2. Limits for
opening and closing speed as well as for the largest and smallest opening of
the control valve are given.
1
1
p
T s
1
s
f '
u
1
G
T
open
u
close
u
min
u
max
u
0
u
1
R
R
T s
T s
G
V
u '
Figure 3.26. Model of turbine governor for hydro turbine.
Parameter Typical Values
T
R
2.5 – 7.5 s
T
G
0.2 – 0.4 s
T
p
0.03 – 0.06 s
δ 0.2 – 1
σ 0.03 – 0.06
Table 3.2. Typical values for some parameters of the turbine governor
for hydro power.
3.5. Dynamic Responses including Turbine Dynamics 47
The controller has two feedback loops, a transient feedback loop and
a static feedback loop. The transient feedback loop has the ampliﬁcation
δ for high frequencies. Thus, the total feedback after a frequency change
is −(δ + σ). In steady state, the transient feedback is zero, and the ratio
between the frequency deviation and the change in the control valve is given
by
∆u = −
1
σ
∆f . (3.52)
Using eq. (3.51), the stationary change of power is obtained as
∆P = −
1
σ
P
0
u
0
∆f . (3.53)
Thus, the speed droop for generator i, S
i
, is
S
i
= σ
i
u
0i
P
0i
, (3.54)
and the total speed droop, S, in the system is given by
S
−1
=
i
S
−1
i
. (3.55)
The transient feedback is needed since the water turbine is a non–minimum
phase system as discussed above. With static feedback only, the control
performance will be unsatisfactory when closedloop stability shall be guar
anteed. Increasing the static feedback to a range where a reasonable control
performance could be attained will make the system unstable.
The transient feedback loop provides an additional feedback component
during nonstationary operating conditions. This feedback decays as steady
state is attained. The initial total feedback can be about ten times larger
than the static feedback, i.e. the speed droop is initially lower than its sta
tionary value.
3.5 Dynamic Responses including Turbine Dynamics
In this section, we will simulate the dynamic responses of a power system
consisting of diﬀerent generation unit types with primary control equipment.
The goal is to improve the understanding of the impact of the turbine dy
namics on the overall control behaviour of the system.
We consider again the onearea power system as presented in Figure 3.11
with the main parameters given according to Table 3.3. As stated before, the
turbine dynamics is inserted in the block ”Turbine dynamics/control”. The
eﬀect of three diﬀerent turbine dynamics will be studied: a steam turbine
without reheater, a steam turbine with reheater and a hydro turbine.
In Figure 3.27, the power system response to a load increase disturbance
of ∆P
load
= +1000 MW is compared. Note that this is equivalent to the
loss of a major generation unit, e.g. a nuclear power plant.
48 3. Primary Frequency Control
Parameter Value Unit
H 5 s
S
B
10 GW
f
0
50 Hz
D
L
1
200
Hz
MW
W
0
0
MW
Hz
T
CH
0.3 s
T
RH
8 s
T
CO
0.5 s
S 0.04 p.u.
T
W
1.4 s
T
G
0.2 s
T
R
6.5 s
σ 0.04 p.u.
δ 0.3 p.u.
Table 3.3. Parameters for time domain simulation of the power system.
The upper plot shows the frequency response of the uncontrolled power
system just as a comparison. It can be seen in the following plots that
the frequency control performance is highly dependent on the turbine type.
While steam turbines without reheating equipment do not allow the fre
quency to decay substantially more than their steadystate deviation, the
time delays caused by the reheater make the response a lot slower. The
hydro turbine allows the most signiﬁcant frequency decay before the system
is brought back to the steadystate frequency.
3.5. Dynamic Responses including Turbine Dynamics 49
0 5 10 15 20 25 30
46
48
50
Uncontrolled power system
f
[
H
z
]
0 5 10 15 20 25 30
49
49.5
50
Primary control on unit without reheating
f
[
H
z
]
0 5 10 15 20 25 30
49
49.5
50
Primary control on unit with reheating
f
[
H
z
]
0 5 10 15 20 25 30
48
49
50
Primary control on hydro unit
f
[
H
z
]
0 5 10 15 20 25 30
0
500
1000
System load step ∆ P
load
∆
P
l
o
a
d
[
M
W
]
Time [s]
Figure 3.27. Dynamic response of the power system without pri
mary control, with primary control on steam turbines without reheater,
steam turbines with reheater, and hydro turbines.
50 3. Primary Frequency Control
Finally, one more illustration of the eﬀect of the turbine dynamics shall
be given. For a direct comparison of a primarycontrolled onearea system
without and with turbine dynamics, the exemplary system simulated in
Figure 3.12 is now simulated again with a steam turbine model without
reheater (T
CH
= 0.3 s). The result is shown in Figure 3.28.
0 1 2 3 4 5 6 7 8 9 10
49.75
49.8
49.85
49.9
49.95
50
S
y
s
t
e
m
F
r
e
q
u
e
n
c
y
[
H
z
]
0 1 2 3 4 5 6 7 8 9 10
0
100
200
300
400
Time [s]
P
o
w
e
r
[
M
W
]
∆ P
load
∆ P
m
∆ P
load
f
Figure 3.28. Behaviour of the onearea system (including dynamics
of steam turbine without reheater, otherwise same parameterization as
in Figure 3.12) after a step increase in load. The upper plot shows the
system frequency f. The lower plot shows the step function in ∆P
load
,
the increase in turbine power ∆P
m
, and the frequencydependent load
variation ∆P
f
load
.
4
Load Frequency Control
In this chapter the secondary, or loadfrequency, control of power systems
will be discussed. Simple models that enable the simulation of the dynamic
behaviour during the action of frequency controllers will also be derived and
studied.
In the previous chapter, the role of the primary frequency control was
dealt with. It was shown that after a disturbance a static frequency error
will persist unless additional control actions are taken. Furthermore, the
primary frequency control might also change the scheduled interchanges be
tween diﬀerent areas in an interconnected system. To restore the frequency
and the scheduled power interchanges, additional control actions must be
taken. This is done through the LoadFrequency Control (LFC). The LFC
can be done either manually through operator interaction or automatically.
In the latter case it is often called Automatic Generation Control (AGC).
The characteristics of AGC will be studied in the subsequent sections, both
during steady state and dynamic conditions.
4.1 Static Characteristics of AGC
The overall purpose of the Automatic Generation Control comprises two
main aspects:
• Keep the frequency in the interconnected power system close to the
nominal value.
1
• Restore the scheduled interchanges between diﬀerent areas, e.g. coun
tries, in an interconnected system.
The mechanisms of AGC that enable it to fulﬁll these requirements will be
outlined in the sequel.
1
In many systems, deviations of up to ±0.1 Hz from the nominal value (50 or 60 Hz)
is deemed as acceptable in steady state. In some systems, North America, even tighter
tolerance bands are applied, while recently in UK the tolerance band has been relaxed
somewhat.
51
52 4. Load Frequency Control
G
G
G
G
G
G G
1 T
P
2 T
P
1
f
2
f
1
1 1
Tie line power for Area 1 Sum over all tie lines =
j
T T
j
P P
eO
= =
_
Area 1
Area 2
AGC
2
AGC
1
2
set
AGC
P 1
set
AGC
P
Figure 4.1. Two area system with AGC.
Consider a two area system as depicted in Figure 4.1. The two secondary
frequency controllers, AGC
1
and AGC
2
, will adjust the power reference
values of the generators participating in the AGC. In an Narea system,
there are N controllers AGC
i
, one for each area i. A block diagram of such
a controller is given in Figure 4.2. A common way is to implement this as a
proportionalintegral (PI) controller:
∆P
AGCi
= −(C
pi
+
1
sT
Ni
)∆e
i
(4.1)
where C
pi
= 0.1 . . . 1.0 and T
Ni
= 30 . . . 200 s. The error ∆e
i
is called
Area Control Error, ACE
i
for area i.
We will now consider a system with N areas. The ACEs are in this case:
ACE
i
=
j∈Ω
i
(P
j
Ti
−P
j
T0i
) +B
i
(f −f
0
) i = 1, 2, . . . , N . (4.2)
Deﬁning now
∆P
Ti
=
j∈Ω
i
(P
j
Ti
−P
j
T0i
) , (4.3)
the ACE can be written as
ACE
i
= ∆P
Ti
+B
i
∆f i = 1, 2, . . . , N . (4.4)
The set Ω
i
consist of all areas connected to area i for which the tie line
powers should be controlled to the set value P
j
T0i
. The constants B
i
are
4.1. Static Characteristics of AGC 53
Ti
P
0 T i
P
0
f
f
AGCi
P '
i
e '
B
i
AGC
i
f '
Figure 4.2. Control structure for AGC (∆e
i
= Error = ACE
i
=
Area Control Error for area i)
called frequency bias factors [MW/Hz]. It is assumed that the frequency
references are the same in all areas, i.e. f
0i
= f
0
for all i, and f is also the
same in steady state for all areas. The goal is to bring all ACE
i
→0.
The variables are thus ∆P
Ti
(N variables) and f, i.e. in total N + 1 vari
ables. Since we have N equations (ACE
i
= 0), we need one more equation.
As a ﬁfth equation we have the power balance:
i
∆P
Ti
= 0 , (4.5)
and consequently a solution can be achieved.
In steady state, f is identical for all areas, and we assume that the
frequency is controlled back to the reference value, i.e. f
0
= f. If the sum
of the reference values of the tie line powers P
j
T0i
is 0, then the system will
settle down to an operating point where P
j
T0i
= P
j
Ti
for all tie line powers.
The time constants of the AGC is chosen such that it reacts much slower
than the primary frequency control.
Selection of Frequency Bias Factors
Consider the two area system in Figure 4.1. The load is now increased with
∆P
load
in area 2. If the tie line power should be kept the same, the generation
must be increased in area 2 with ∆P
load
after the AGC has reacted. Before
the AGC has reacted we have a frequency deviation of ∆f in both areas,
54 4. Load Frequency Control
which means that for area 1
∆f = −S
1
∆P
T1
(4.6)
and for area 2
∆f = −S
2
(∆P
load
−∆P
T1
) (4.7)
and ∆P
T1
= −∆P
T2
. The two ACEs can now be written as
ACE
1
= ∆P
T1
+B
1
∆f = ∆P
T1
+B
1
(−S
1
∆P
T1
) = ∆P
T1
(1 −B
1
S
1
) (4.8)
and
ACE
2
= ∆P
T2
+B
2
∆f = ∆P
T2
+B
2
(−S
2
(∆P
load
−∆P
T1
)) =
∆P
T2
(1 −B
2
S
2
) −B
2
S
2
∆P
load
(4.9)
In this case it is desirable that the AGC controller in area 1 does not react.
If we set B
1
= 1/S
1
we see from eq. (4.8) that ACE
1
= 0. This is called
Non Interactive Control. If B
2
= 1/S
2
is chosen the ACE in area 2 becomes
ACE
2
= −∆P
load
(4.10)
This means that only controller 2 reacts and the load increase ∆P
load
is
compensated for in area 2 by the PI control law as stated in eq. (4.1).
However, as long as the controller in eq. (4.1) has an integrating part,
all positive values of B
i
will guarantee that all ACE
i
→ 0. The choice ac
cording to Non Interactive Control has been found to give the best dynamic
performance through a number of investigations. In a multiarea case this
corresponds to selecting B
i
= 1/S
i
for all areas.
4.2 Dynamic Characteristics of AGC
As in the previous chapter on primary frequency control, we will now develop
a dynamic model of the power system where the newly introduced AGC is
included. As the way of extracting inputoutput transfer functions from
block diagrams has already been presented in the last chapter, we omit the
analytical derivation here for shortness. The interested reader is welcome
to calculate the transfer function between the load disturbance and the
frequency as an exercise.
4.2.1 Onearea system
In a singlearea power system, there is obviously no tieline power to be
controlled. In this case, the AGC only fulﬁlls the purpose of restoring the
nominal system frequency. Figure 4.3 shows the known block diagram which
has been extended by the proportionalintegral control law of the AGC.
Figure 4.4 shows a plot of the system frequency which is brought back to
the nominal value. The parameters of the system are as presented in Table
3.3 (turbine dynamics neglected), and B = 1/S, C
p
= 0.17, T
N
= 120 s.
4.2. Dynamic Characteristics of AGC 55
load
P A
f A
÷
e
P A
+
+
+
f
load
P A
+
+
+
+
+
+
0
0
2W
s
f
Rotating mass loads
Frequencydependent loads
System inertia
System load change
1
l
D
0
(2 )
B
f
HS s
m
P A
1
1
t
T s +
set
m
P A
0
set
m
P A
1
S
÷
Turbine
dynamics/
control
Primary
control
1
) (
p
N
C
s T
÷ +
ACE
AGC
P A
sched
P A
AGC
(one area
system)
B
frequency
bias
Figure 4.3. Dynamic model of onearea system with AGC.
0 100 200 300 400 500 600 700 800 900
49.7
49.75
49.8
49.85
49.9
49.95
50
50.05
50.1
Time [s]
f
[
H
z
]
with ACG
without AGC
Figure 4.4. Dynamic response of the onearea system equipped with
primary control and AGC compared with the same system without
AGC.
56 4. Load Frequency Control
4.2.2 Twoarea system – unequal sizes – disturbance response
Here we consider the same twoarea power system as in section 3.3.4 where
Area 1 is assumed to be much smaller than Area 2. The corresponding block
diagram including the AGC is shown in Figure 4.6. Because of the secondary
control (AGC) one obtains the step response in Figure 4.5. The load increase
of ∆P
load,1
= 1000 MW in Area 1 is in this case fully compensated by the
generators in Area 1.
0 100 200 300 400 500 600
49.8
49.85
49.9
49.95
50
50.05
S
y
s
t
e
m
f
r
e
q
u
e
n
c
y
[
H
z
]
f
1
f
2
0 100 200 300 400 500 600
−200
0
200
400
600
800
1000
Time [s]
P
[
M
W
]
∆ P
AGC,1
∆ P
AGC,2
∆ P
T21
Figure 4.5. Step response for the system in Figure 4.6 with AGC. The
upper diagram shows the system frequencies of Area 1 and Area 2. The
lower diagram shows the control action of the AGCs in Area 1 ∆P
AGC,1
which compensates the power deﬁcit and in Area 2 ∆P
AGC,2
which
is largely uninﬂuenced by the disturbance. Furthermore, the tie line
power ∆P
T21
is shown, which initially compensates the disturbance
and then slowly decays to 0.
4.2. Dynamic Characteristics of AGC 57
2
B
,1
1
l
D
,1 load
P A
1
1
1
( )
p
N
C
s T
÷ +
2
2
1
( )
p
N
C
s T
÷ +
ˆ
2
T
P
s
t
,2 sched
P A
,2 AGC
P A
12 T
P A
12 T
P A
1
1
t
T s +
1
1
t
T s +
+
+
+
÷ ÷
÷
+
+ +
+ +
Figure 4.6. Dynamic model of twoarea system with AGC.
58 4. Load Frequency Control
4.2.3 Twoarea system, unequal sizes – normal control operation
So far, we only simulated the dynamic response of the power system to
stepwise disturbances in the load. During the everyday normal operation
of a power system, however, the automatic control loops of primary and
secondary control are always active to compensate for the continuously aris
ing small imbalances between generation and load. These normaloperation
imbalances are due to load ﬂuctuations, load prediction errors, and, in the
presence of intermittent infeeds (from wind power etc.) in the control area,
infeed ﬂuctuations and infeed prediction errors. In this context, the diﬀer
ence between ﬂuctuation and prediction errors is constituted by the time
scales on which they occur: while the term ”ﬂuctuation” is usually deﬁned
as largely uncorrelated stochastic noise on a second or minute time scale,
”prediction error” usually refers to a mismatch of predicted and actual con
sumption/infeed on a 15minute time scale. The latter usually shows a
certain autocorrelation, which means that the probability of the signal to
assume a certain value is dependent on the previously assumed signal values.
Another important impact on the AGC arises from schedule changes
that occur at the full hour. The reason for this timing is the trading on
electricity markets which is usually done in an hourly framework. Because
of ramping actions of power plants which often deviate from the ramping
proﬁles they are supposed to follow, imbalances just before the full hour
(either positive or negative), as well as just after the full hour (then in the
opposite direction) arise. This eﬀect, however, is disregarded in the present
setup.
Figure 4.7 demonstrates the response of the twoarea system to a stochas
tically varying load. The system is parameterized again according to Table
3.1. To increase the degree of realism, the dynamic model of a steam tur
bine without reheater is included. As often found in practice, the primary
controls are equipped with a dead band of ±10 mHz. Saturation of the
control variables (due to limited amounts of contracted control reserves) is
not modelled. A detailed discussion of the synthesis of the load ﬂuctuation
and prediction error signals is beyond the scope of this lecture. Thus, only
the input time series ∆P
load,1
and ∆P
load,2
are shown in the plot.
4.2. Dynamic Characteristics of AGC 59
0 10 20 30 40 50 60
49.95
50
50.05
50.1
50.15
System frequencies
S
y
s
t
e
m
f
r
e
q
u
e
n
c
y
[
H
z
]
f
1
f
2
0 10 20 30 40 50 60
−400
−200
0
200
400
600
Activity of secondary control
P
[
M
W
]
∆ P
AGC,1
∆ P
AGC,2
/1000
∆ P
T21
0 10 20 30 40 50 60
−400
−200
0
200
400
Activity of primary control
P
[
M
W
]
∆ P
m,1
set
∆ P
m,2
set
/1000
0 10 20 30 40 50 60
−400
−200
0
200
400
Stochastic load fluctuation and prediction error
Time [min]
P
[
M
W
]
∆ P
load,1
∆ P
load,2
/ 1000
Figure 4.7. Normal operation of the two area system. The usual
small mismatches between generation and demand are modelled by a
stochastic disturbance in the load.
60 4. Load Frequency Control
4.2.4 Twoarea system – equal sizes, including saturations – dis
turbance response
Finally we brieﬂy investigate the dynamic behaviour of a twoarea system
where both areas are of the same size. Turbine dynamics are again neglected
in this scenario. Here, saturations of the control inputs which represent the
limited amount of contracted primary and secondary control reserves are
modelled (primary control: ±75 MW, secondary control: ±350 MW). A
graphical representation of the saturations in a block diagram is omitted
here for shortness. Referring again to Figure 4.6, the saturation blocks
would have to be inserted in the signal path of ∆P
set
m
(primary control)
and ∆P
AGC
(secondary control). Furthermore, countermeasures against
integrator windup have to be taken, which shall not be discussed here.
Figure 4.8 shows the step response (∆P
load,1
= 400 MW) for the system
with two equally sized areas. It is clearly visible that the AGC of Area 1 can
not compensate for the deviation and remains at its saturation of +350 MW.
The system frequency is not brought back to 50 Hz. In real power system
operation, the ancillary service dispatcher in the control center would now
manually activate tertiary control reserves in order to relieve the saturated
secondary reserves.
0 100 200 300 400 500 600
49.88
49.9
49.92
49.94
49.96
49.98
50
f
[
H
z
]
∆ f
1
∆ f
2
0 100 200 300 400 500 600
−100
0
100
200
300
400
Time [s]
P
[
M
W
]
∆ P
T12
∆ P
AGC,1
∆ P
AGC,2
Figure 4.8. Step response for the system in Figure 4.6 with AGC,
equal areas with a rated power of 10 GW each. Because of the satu
rated secondary control, the power system frequency cannot be brought
back to the nominal value of 50 Hz.
5
Synchronous Machine Model
Almost all energy consumed by various loads in an electric power system is
produced by synchronous machines, or, more correctly, the conversion from
the primary energy sources, like water energy, nuclear energy, or chemical
energy, to electrical energy is done in synchronous machines with a mechan
ical intermediate link, the turbine. This is true in larger power systems, but
not always in smaller systems like isolated islands, power supply of equip
ment in deserts, or other smaller systems. In these systems, the energy
can come from asynchronous generators, for example in wind generation
units, batteries, or some other source of electrical energy. In systems with
synchronous generators, these have an extremely important part in many
dynamic phenomena. Thus, it is very important to develop usable and real
istic models of the synchronous machines. In the previous chapters, mainly
the mechanical properties of the synchronous machines have been modelled
using the swing equation, while a very simplistic model of the electrical prop
erties of the synchronous machine has been used. In this chapter, a more
general, detailed model of the electric parts of the synchronous machine will
be derived. The simple models used earlier will be justiﬁed. It should be
emphasized that the description here aims towards the development of mod
els usable for studying dynamic phenomena in the power system. It is not
the purpose of these models to give a detailed and deep understanding of the
physical functions of the synchronous machine. Of course, it is desirable to
have a good insight into the physics of the synchronous machine to be able to
derive appropriate models. For a detailed discussion of these aspects, books
and courses dealing with the theory of electrical machines should be studied.
5.1 Park’s Transformation
Park’s transformation is a phase transformation (coordinate transformation)
between the three physical phases in a three phase system and three new
phases, or coordinates, that are convenient for the analysis of synchronous
machines. This transformation is also known as the dq–transformation or
Blondel’s transformation. A reason why the transformation is suitable can
be derived from Figure 5.1.
It is obvious that the phase quantities in the a–, b–, and c–phases will
vary periodically in steady state. Further, the self and mutual inductances
between stator circuits and rotor circuits will vary with the rotor position.
Instead of performing all computations in the ﬁxed stator system, the stator
61
62 5. Synchronous Machine Model
Direction
of Rotation
n′
i
a
fa sa
sc
sb
i
Q
θ
qaxis
aaxis
daxis
baxis
caxis
i
Q
i
D
i
D
i
F
i
F
i
b
fb
n′
n′
fc
i
c
Figure 5.1. Deﬁnition of quantities in Park’s transformation.
quantities voltages, currents, and ﬂuxes can be transformed to a system that
rotates with the rotor. Thus, two orthogonal axes are deﬁned as shown in
Figure 5.1: One along the axis in which the current in the rotor windings
generates a ﬂux, and one in an axis perpendicular to this. The ﬁrst is the
direct axis (d–axis), and the other is the quadrature axis (q–axis). From now
on, the denominations d–axis and q–axis will be used. To make the system
complete, a third component corresponding to the zero sequence must be
deﬁned.
Figure 5.1 is a simpliﬁed picture of a synchronous machine and should
only be viewed as an intuitive basis for the transformation given below. The
machine in Figure 5.1 has one pole pair, but Park’s transformation can, of
course, be applied to machines with an arbitrary number of pole pairs.
Park’s transformation is, as a consequence of the reasoning above, time
dependent, and the connection between the phase currents and the trans
5.1. Park’s Transformation 63
formed currents is given by
_
¸
¸
¸
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
¸
¸
¸
_
i
0
=
_
1
3
(i
a
+i
b
+i
c
) ,
i
d
=
_
2
3
_
i
a
cos θ +i
b
cos
_
θ −
2π
3
_
+i
c
cos
_
θ +
2π
3
__
,
i
q
=
_
2
3
_
−i
a
sin θ −i
b
sin
_
θ −
2π
3
_
−i
c
sin
_
θ +
2π
3
__
.
(5.1)
If the a–axis is chosen as reference,
θ = ωt +θ
0
(5.2)
is obtained and the time dependence in the transformation is obvious. It
should be pointed out that i
a
, i
b
, and i
c
are the real physical phase currents
as functions of time and not a phasor representation of those. Now,
x
abc
= (x
a
, x
b
, x
c
)
T
x
0dq
= (x
0
, x
d
, x
q
)
T
(5.3)
can be deﬁned. x can here be an arbitrary quantity, like voltage, current,
or ﬂux. With this notation, Park’s transformation can be written as
x
0dq
= Px
abc
(5.4)
with
P =
_
2
3
_
_
1/
√
2 1/
√
2 1/
√
2
cos θ cos(θ −
2π
3
) cos(θ +
2π
3
)
−sinθ −sin(θ −
2π
3
) −sin(θ +
2π
3
)
_
_
. (5.5)
The inverse transformation is then given by
x
abc
= P
−1
x
0dq
, (5.6)
and it can easily be shown that
P
−1
= P
T
. (5.7)
A mnemonic for Park’s transformation can be obtained from Figure 5.1 by
projecting the a–, b–, and c–axes onto the d– and q–axes in the ﬁgure.
Equation (5.7) implies that Park’s transformation is an orthonormal
transformation. This is reﬂected in the expression for the momentary power
that is produced in the stator windings
p = u
a
i
a
+u
b
i
b
+u
c
i
c
= u
T
abc
i
abc
=
(P
−1
u
0dq
)
T
P
−1
i
0dq
= (P
T
u
0dq
)
T
P
−1
i
0dq
=
u
T
0dq
PP
−1
i
0dq
= u
T
0dq
i
0dq
=
u
0
i
0
+u
d
i
d
+u
q
i
q
.
(5.8)
64 5. Synchronous Machine Model
Here, Equations (5.6) and (5.7) have been used. Equation (5.8) can therefore
be written as
p = u
a
i
a
+u
b
i
b
+u
c
i
c
= u
0
i
0
+u
d
i
d
+u
q
i
q
. (5.9)
Equation (5.9) shows that the introduced transformation is power invariant,
which is a consequence of Equation (5.7).
It should be pointed out that there are several diﬀerent variants of Park’s
transformation appearing in literature. They can diﬀer from the form pre
sented here by the direction of the q–axis and by constants in the transfor
mation matrix. When using equations from some book or paper, it is thus
important to make sure that the deﬁnition of Park’s transformation used is
the same as one’s own. Otherwise, wrong results might be obtained.
The rotor windings produce a ﬂux linkage that mainly lies in the direction
of the d–axis. That ﬂux induces an electromagnetic ﬁeld, E, which is
lagging by 90
◦
, hence in the direction of the negative q–axis. For generator
operation, the phasor for E leads by an angle δ before the phasor for the
terminal voltage U. At t = 0, the negative q–axis thus leads by an angle δ
before the phasor for the voltage along the a–axis, cf. Figure 5.1. For t > 0,
the d– and q–axes have moved by an angle ωt with the angular speed of the
rotor ω. The rotor’s d–axis will hence be in position
θ = ωt +δ +
π
2
. (5.10)
It is particularly of interest to study how zero sequence, negative sequence,
and positive sequence quantities are transformed by Park’s transformation.
It is comparatively easy to show that a pure zero sequence quantity only
leads to a contribution in x
0
with x
d
= x
q
= 0. A pure positive sequence
quantity
x
abc
(+) =
√
2x
_
_
sin(θ +α)
sin(θ +α −
2π
3
)
sin(θ +α +
2π
3
)
_
_
(5.11)
is transformed to
x
0dq
(+) =
√
3x
_
_
0
sin(α)
−cos(α)
_
_
, (5.12)
i.e. pure DCquantities (time independent) in the dq–system with the zero
sequence component equal zero. A pure negative sequence quantity gives
rise to quantities in d– and q–directions that vary with the angular frequency
2ω. The zero sequence component vanishes also in this case. (Show this!)
5.2. The Inductance Matrices of the Synchronous Machine 65
5.2 The Inductance Matrices of the Synchronous Ma
chine
In the following, a synchronous machine with one damper winding in d– and
one in q–axis and, of course, a ﬁeld winding is considered. Quantities related
to these windings are denoted with the indices D, Q, and F, respectively.
The ﬂux linkages in the stator windings (Ψ
abc
) and in the windings F, D,
and Q (Ψ
FDQ
) depend on the currents in these windings according to
_
Ψ
abc
Ψ
FDQ
_
=
_
L
abc,abc
L
abc,FDQ
L
FDQ,abc
L
FDQ,FDQ
__
i
abc
i
FDQ
_
. (5.13)
L
abc,abc
, . . ., L
FDQ,FDQ
are 3 ×3 matrices with self and mutual inductances
as matrix elements. These will depend on the rotor position, hence they are
time dependent. The following approximation for inductances L(θ) proves
to be useful: L(θ) = L
s
+ L
m
cos 2θ with L
s
> L
m
> 0, where L
s
and L
m
are machine constants that can be identiﬁed from measurements. The same
holds for the machine constant M
s
that will be used in the case of mutual
inductances. Keeping that in mind, the matrix elements are calculated.
• L
abc,abc
:
_
¸
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
¸
_
L
aa
= L
s
+L
m
cos 2θ ,
L
bb
= L
s
+L
m
cos(2θ −
4π
3
) ,
L
cc
= L
s
+L
m
cos(2θ +
4π
3
) ,
L
ab
= L
ba
= −M
s
−L
m
cos(2θ +
π
3
) ,
L
bc
= L
cb
= −M
s
−L
m
cos(2θ +π) ,
L
ac
= L
ca
= −M
s
−L
m
cos(2θ +
5π
3
) .
(5.14)
• L
abc,FDQ
and
L
FDQ,abc
:
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
L
aF
= L
Fa
= M
F
cos θ ,
L
bF
= L
Fb
= M
F
cos(θ −
2π
3
) ,
L
cF
= L
Fc
= M
F
cos(θ +
2π
3
) ,
L
aD
= L
Da
= M
D
cos θ ,
L
bD
= L
Db
= M
D
cos(θ −
2π
3
) ,
L
cD
= L
Dc
= M
D
cos(θ +
2π
3
) ,
L
aQ
= L
Qa
= −M
Q
sin θ ,
L
bQ
= L
Qb
= −M
Q
sin(θ −
2π
3
) ,
L
cQ
= L
Qc
= −M
Q
sin(θ +
2π
3
) .
(5.15)
• L
FDQ,FDQ
:
_
¸
¸
¸
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
¸
¸
¸
_
L
FF
= L
F
,
L
DD
= L
D
,
L
QQ
= L
Q
,
L
FD
= L
DF
= M
R
,
L
FQ
= L
QF
= 0 ,
L
DQ
= L
QD
= 0 .
(5.16)
66 5. Synchronous Machine Model
All inductances with only one index in Equations (5.14) – (5.16) are con
stants and depend on the design of the synchronous machine. The resulting
inductances are of course, as mentioned before, not quite exact. They can be
called exact in an ideal machine, where spatial harmonics and other unsym
metries are neglected. For a real synchronous machine, the approximations
are usually very good and lead to fully acceptable results for the computa
tions and analyses treated here. It should be emphasized that the model
developed here is for use in computations where the synchronous machines
are part of a larger system. The model is not primarily aimed at studies of
the internal quantities in the generator.
It is now natural to transform the abc–components in Equation (5.13)
to 0dq–components. For this, an extended transformation given by
P
ex
=
_
P 0
0 I
_
, (5.17)
with P according to (5.5) and a 3 ×3 unit matrix I is used. The result is
_
Ψ
0dq
Ψ
FDQ
_
=
_
L
0dq,0dq
L
0dq,FDQ
L
FDQ,0dq
L
FDQ,FDQ
__
i
0dq
i
FDQ
_
, (5.18)
with the inductance matrix given by
L
0dq,0dq
L
0dq,FDQ
L
FDQ,0dq
LFDQ,FDQ
=
P 0
0 I
L
abc,abc
L
abc,FDQ
L
FDQ,abc
LFDQ,FDQ
P
−1
0
0 I
.
(5.19)
The virtue of the Park’s transformation is apparent in the following equa
tion, where the inductance matrix in (5.18) is computed
L
0dq,0dq
=
_
_
L
0
0 0
0 L
d
0
0 0 L
q
_
_
, (5.20)
with
_
¸
_
¸
_
L
0
= L
s
−2M
s
,
L
d
= L
s
+M
s
+
3
2
L
m
,
L
q
= L
s
+M
s
−
3
2
L
m
,
(5.21)
and
L
0dq,FDQ
=
_
_
0 0 0
kM
F
kM
D
0
0 0 kM
Q
_
_
, (5.22)
where
k =
_
3
2
(5.23)
and, of course,
L
FDQ,0dq
= L
T
0dq,FDQ
. (5.24)
5.3. Voltage Equations for the Synchronous Machine 67
ω
d
daxis
qaxis
q
Q
F
D
Figure 5.2. Schematic picture of the transformed system. (One
damper winding in the d–axis and one in the q–axis.)
L
FDQ,FDQ
has, of course, not changed, but for completeness it is re
peated here.
L
FDQ,FDQ
=
_
_
L
F
M
R
0
M
R
L
D
0
0 0 L
Q
_
_
. (5.25)
Two important observations can be made from Equations (5.20)–(5.25):
• The inductances in the inductance matrix in Equation (3.18) are not
dependent on time.
• The quantities in d and qdirections are decoupled. (The induction
matrix is block diagonal: one 2 ×2 matrix and one 1 ×1 matrix.)
The second observation above leads to a picture of the transformed sys
tem according to Figure 5.2.
5.3 Voltage Equations for the Synchronous Machine
For the three stator circuits and the three rotor circuits the following rela
tions can be written:
_
_
_
u
a
= −r
a
i
a
−
˙
Ψ
a
,
u
b
= −r
b
i
b
−
˙
Ψ
b
,
u
c
= −r
c
i
c
−
˙
Ψ
c
,
(5.26)
68 5. Synchronous Machine Model
Figure 5.3. Sign convention for stator and rotor circuits.
and
_
_
_
u
F
= r
F
i
F
+
˙
Ψ
F
,
0 = r
D
i
D
+
˙
Ψ
D
,
0 = r
Q
i
Q
+
˙
Ψ
Q
.
(5.27)
Equations (5.26) and (5.27) can be written more compactly in vector
form,
_
u
abc
u
FDQ
_
= −
_
R
abc
0
0 R
FDQ
__
i
abc
i
FDQ
_
−
_
˙
Ψ
abc
˙
Ψ
FDQ
_
. (5.28)
The vector u
FDQ
is deﬁned as
u
FDQ
= (−u
F
, 0, 0)
T
, (5.29)
while the other vectors are deﬁned as before. R
abc
and R
FDQ
are diagonal
3 ×3 matrices.
If Equation (5.28) is multiplied by P
ex
according to Equation (5.17), all
quantities are transformed to the dq–system, i.e.
_
u
0dq
u
FDQ
_
= −
_
PR
abc
P
−1
0
0 R
FDQ
__
i
0dq
i
FDQ
_
−
_
P
˙
Ψ
abc
˙
Ψ
FDQ
_
. (5.30)
The matrix PR
abc
P
−1
is denoted R
0dq
, and if r
a
= r
b
= r
c
= r, which in
most cases is true,
R
0dq
= R
abc
= rI (5.31)
is valid.
5.3. Voltage Equations for the Synchronous Machine 69
To get Equation (5.30) expressed solely in dq–quantities, also the last
term on the right hand side must be expressed in these. Since P is time
dependent, it is important to remember that
˙
P = 0 , (5.32)
which leads to
˙
Ψ
0dq
=
d
dt
(PΨ
abc
) =
˙
PΨ
abc
+P
˙
Ψ
abc
, (5.33)
and thus
P
˙
Ψ
abc
=
˙
Ψ
0dq
−
˙
PΨ
abc
=
˙
Ψ
0dq
−
˙
PP
−1
Ψ
0dq
. (5.34)
Equation (5.30) can hence be written as
_
u
0dq
u
FDQ
_
= −
_
R
0dq
0
0 R
FDQ
__
i
0dq
i
FDQ
_
−
_
˙
Ψ
0dq
˙
Ψ
FDQ
_
+
_
˙
PP
−1
Ψ
0dq
0
_
.
(5.35)
Some trivial computations show that the matrix
˙
PP
−1
can be expressed as
˙
PP
−1
=
_
_
0 0 0
0 0 ω
0 −ω 0
_
_
. (5.36)
The voltage equations in the dq–system can thus be written in component
form as
_
_
_
u
0
= −ri
0
−
˙
Ψ
0
,
u
d
= −ri
d
−
˙
Ψ
d
+ωΨ
q
,
u
q
= −ri
q
−
˙
Ψ
q
−ωΨ
d
,
(5.37)
and
_
_
_
−u
F
= −r
F
i
F
−
˙
Ψ
F
,
0 = −r
D
i
D
−
˙
Ψ
D
,
0 = −r
Q
i
Q
−
˙
Ψ
Q
.
(5.38)
In the previous section, expressions for the dependencies of the ﬂux linkages
on the currents in the diﬀerent windings were derived. To further simplify
the expressions that were obtained, the per unit system for the diﬀerent
windings is now introduced so that all mutual inductances in the d–axis
are equal, and all in the q–axis are equal. (In our case, only one damping
winding in the q–axis was considered, but in a more general case several
damper windings can be considered.) We introduce
_
3
2
M
F
=
_
3
2
M
D
= M
R
= L
AD
,
_
3
2
M
Q
= L
AQ
.
(5.39)
70 5. Synchronous Machine Model
The diﬀerent ﬂuxes can then be written as
_
_
_
Ψ
0
= L
0
i
0
,
Ψ
d
= L
d
i
d
+L
AD
(i
F
+i
D
) ,
Ψ
q
= L
q
i
q
+L
AQ
i
Q
,
(5.40)
and
_
_
_
Ψ
F
= L
F
i
F
+L
AD
(i
d
+i
D
) ,
Ψ
D
= L
D
i
D
+L
AD
(i
d
+i
F
) ,
Ψ
Q
= L
Q
i
Q
+L
AQ
i
q
.
(5.41)
Equations (5.37) and (5.38) together with Equations (5.40) and (5.41) now
describe the electrical dynamics of a synchronous machine completely. These
equations together with a description of the external system unequivocally
determine the behaviour of the synchronous machine during diﬀerent dis
turbances. In Figure 5.4, a graphical description of these equations is given.
In Equation (5.37), we observe that the emf in d– and q–direction consists
of two terms: one that is a time derivative of the absolute value of the ﬂux
linkage and one that arises because the ﬁeld winding is rotating. The ﬁrst
of these is usually called stator transient and the other rotational emf. In
steady state, the ﬁrst of these vanishes, and the whole emf is created by the
rotation of the ﬁeld winding. It can be shown that the terms
˙
Ψ
d
and
˙
Ψ
q
are
in most applications much smaller than ωΨ
d
and ωΨ
q
, which justiﬁes that
the ﬁrst ones are often neglected.
5.4 Synchronous, Transient, and Subtransient Induc
tances
The complete description of the synchronous machine given in the previous
section can be simpliﬁed and made more understandable from a physical
standpoint if a number of new parameters are introduced. These new pa
rameters can be derived from the already deﬁned ones. In steady state or a
suﬃciently long time after a disturbance, the induced currents in the damper
windings and in the ﬁeld winding vanish. After a disturbance diﬀerent time
intervals can be distinguished, during which diﬀerent couplings between the
circuits in Figure 5.4 prevail. This makes it possible to deﬁne new induc
tances describing the synchronous machine during these time periods, and
these inductances will be derived in this section. Further, the time con
stants that specify how fast the currents in the damper windings decay are
derived in the next section. These time constants indicate also the duration
of the time intervals, during which the derived inductances determine the
behaviour of the synchronous machine.
Earlier, the synchronous inductances L
d
and L
q
were deﬁned. They are
5.4. Synchronous, Transient, and Subtransient Inductances 71
L
d
r
d i
d
+
u
d
L
F
r
F
i
F
u
F
+

+

L
q
r
q i
q
+
u
q
+ 
L
0
r
0 i
0
+
u
0
L
D
r
D
i
D
L
Q
r
Q
i
Q
ωψ
d
ωψ
q
L
AD
L
AQ
Rotor Circuits Stator Circuits
Figure 5.4. Graphical description of the voltage equations and the
coupling between the equivalent circuits.
repeated here for completeness.
_
L
d
= L
s
+M
s
+
3
2
L
m
L
q
= L
s
+M
s
−
3
2
L
m
(5.42)
For a synchronous machine with salient poles, like a hydro power generation
unit, L
m
> 0 and thus L
d
> L
q
, while L
m
≈ 0 for machines with round
rotors leading to L
d
≈ L
q
. These inductances describe the synchronous
machine in steady state as can be seen from Figure 5.4. In steady state the
currents i
D
and i
Q
= 0, so there is no inﬂuence from the damper windings
on the stator circuits. Since in steady state the rotor current i
F
is a pure
dc current the only inﬂuence from the ﬁeld current on the stator circuits
is through the rotational voltages ωΨ
d
and ωΨ
q
. From Figure 5.4 it is
thus obvious that the inductances seen from the generator terminals are in
72 5. Synchronous Machine Model
steady state L
d
and L
q
. However, when transients occur, e.g. due to short
circuits or other disturbances, there will be couplings between the circuits
in Figure 5.4 and this will result in that the equivalent inductances as seen
from the terminals will change. These equivalent inductances will be derived
in the following.
First, the subtransient and transient inductances in the daxis will be
derived. It is assumed that the machine is in steady state before the distur
bance and a voltage change of the form
_
_
∆u
a
∆u
b
∆u
c
_
_
=
√
2∆u
_
_
sin(θ +α)
sin(θ +α −
2π
3
)
sin(θ +α −
4π
3
)
_
_
c(t) (5.43)
is superimposed on the terminal voltages. The function c(t) is a step function
at t = 0, i.e. c(t) = 0 for t < 0 and c(t) = 1 for t > 0. With α = π/2 the
Park–transformed voltage vector in Equation (5.43) becomes
_
_
∆u
0
∆u
d
∆u
q
_
_
=
√
3∆u
_
_
0
c(t)
0
_
_
(5.44)
For t = 0+, that is, directly after the voltage is applied on the terminals,
the ﬂux linkages Ψ
F
and Ψ
D
are still zero since they cannot change instan
taneously. This gives the following equations
_
Ψ
F
(0+) = 0 = L
F
∆i
F
+L
AD
(∆i
d
+ ∆i
D
)
Ψ
D
(0+) = 0 = L
D
∆i
D
+L
AD
(∆i
d
+ ∆i
F
)
(5.45)
where ∆i
f
, ∆i
d
, and ∆i
D
, with obvious notation, are the currents induced
in the circuits due to the step in the voltage. From (5.45) ∆i
D
and ∆i
F
can
be expressed in ∆i
d
,
_
¸
¸
¸
¸
_
¸
¸
¸
¸
_
∆i
D
= −
L
F
L
AD
−L
2
AD
L
F
L
D
−L
2
AD
∆i
d
,
∆i
F
= −
L
D
L
AD
−L
2
AD
L
F
L
D
−L
2
AD
∆i
d
.
(5.46)
These expressions for ∆i
D
and ∆i
F
can now be inserted into the expression
for ∆Ψ
d
in (5.40) giving
∆Ψ
d
= L
d
∆i
d
+L
AD
(∆i
F
+ ∆i
D
) =
_
L
d
−
L
D
L
2
AD
+L
F
L
2
AD
−2L
3
AD
L
F
L
D
−L
2
AD
_
∆i
d
= L
′′
d
∆i
d
,
. (5.47)
where the subtransient inductance L
′′
d
in the d–axis has been deﬁned and it
can be rewritten as
L
′′
d
= L
d
−
L
D
+L
F
−2L
AD
L
F
L
D
/L
2
AD
−1
. (5.48)
5.4. Synchronous, Transient, and Subtransient Inductances 73
The time constant of the decay of the current in the damper winding is
much smaller that the time constant in the ﬁeld winding, see next section,
which means that after a certain time it can be assumed that the current
in the damper winding has decayed to zero, i.e. ∆i
D
= i
D
= 0. The same
assumption regarding i
D
can be made if no damper winding is modelled in
the daxis. This assumption gives
∆Ψ
F
= 0 = L
F
∆i
F
+L
AD
∆i
d
(5.49)
which gives
∆i
F
= −(L
AD
/L
F
)∆i
d
(5.50)
and with use of (5.40) and i
D
= 0
∆Ψ
d
=
_
L
d
−
L
2
AD
L
F
_
∆i
d
= L
′
d
i
d
, (5.51)
The transient inductance in the d–axis, L
′
d
, is deﬁned in Equation (5.51) as
L
′
d
= L
d
−
L
2
AD
L
F
. (5.52)
The time constant for the decay of the current in the damper winding is
derived in the next section.
An equivalent analysis to the one above can be performed for the q–
axis, but, since no ﬁeld winding exists in the q–axis for the model we are
considering here, the terminology is somewhat diﬀerent. For a machine with
salient poles and damper winding in the q–axis, the eﬀective inductance after
the current in the damper winding has decayed is practically equal to the
synchronous inductance. Hence it is sometimes said that, for machines with
salient poles, the transient and synchronous inductances in the q–axis are
equal.
With α = π the Park–transformed voltage vector in Equation (5.43)
becomes
_
_
∆u
0
∆u
d
∆u
q
_
_
=
√
3∆u
_
_
0
0
c(t)
_
_
(5.53)
A similar analysis as above will the give the subtransient inductance in q–
axis
L
′′
q
= L
q
−
L
2
AQ
L
Q
(5.54)
for a synchronous machine with one damper winding in the q–axis. Accord
ing to the reasoning above we have
L
′
q
= L
q
. (5.55)
74 5. Synchronous Machine Model
in this case.
If one additional damper winding is modelled into the q–axis, a value of
L
′
q
can be derived that diﬀers from L
q
. This derivation is analogous to the
one done for the inductances in the d–axis.
In general, for the inductances deﬁned, the following relations hold
L
′′
d
< L
′
d
< L
d
,
L
′′
q
< L
′
q
< L
q
.
(5.56)
The subtransient, transient, and synchronous reactances are deﬁned in an
obvious way, e.g. X
d
= ω
0
L
d
etc., which most often are expressed in p.u.
based on the machine power and voltage ratings, see Table 5.1.
5.5 Time constants
The time constants that determine how fast transients in the rotor and
damper windings decay after a disturbance depend on the resistances of
these windings and associated inductances. We will here derive the so called
open circuit time constants whereby it assumed that the stator circuits are
open, i.e. i
a
= i
b
= i
c
= 0, or equivalently i
0
= i
d
= i
q
= 0. These
time constants are normally denoted by an index o for open, e.g. T
′′
do
as
explained below. An alternative that also occurs in the literature are the
time constants when the stator terminals are shortcircuited, which also
uniquley determine the winding resistances. The open circuit time constants
are normally the ones given by the manufacturer of the generator and can
be obtained by measurements.
Thus, suppose now that the stator windings are open and and a step
change in the ﬁeld voltage at the time t = 0, i.e. u
F
= u
F
(0) + ∆u
F
c(t)
will be investigated, with c(t) being a step function at t = 0. We will here
derive the time constants associated with the ﬁeld winding and the damper
winding in the d–axis. The circuit equations describing the currents induced
by the step in the ﬁeld voltage for these two windings are for t > 0
_
r
F
∆i
F
+ ∆
˙
Ψ
F
= ∆u
F
c(t)
r
D
∆i
D
+ ∆
˙
Ψ
D
= 0
(5.57)
For a synchronous machine r
D
≫ r
F
and the induced current will conse
quently decay much faster in the damper winding than in the ﬁeld winding.
(This fact was also used above in the derivation of L
′′
d
and L
′
d
.) Therefore
we will analyze the decay of the current in the damper winding ﬁrst. Since
i
d
= 0, the ﬂuxes in the ﬁeld and damper windings can be written as
_
∆Ψ
D
= L
D
∆i
D
+L
AD
∆i
F
,
∆Ψ
F
= L
F
∆i
F
+L
AD
∆i
D
.
(5.58)
5.5. Time constants 75
Taking the time derivative of the lower equation in (5.58) gives
∆
˙
Ψ
F
= L
F
∆
˙
i
F
+L
AD
∆
˙
i
D
. (5.59)
From the upper equation of (5.57) we can replace ∆
˙
Ψ
F
in (5.59), which after
some rearrangements gives
∆
˙
i
F
=
∆u
F
−r
F
∆i
F
−L
AD
∆
˙
i
D
L
F
(5.60)
Now the time derivative of the upper equation of (5.58) is inserted into the
lower equation of (5.57) yielding
r
D
∆i
D
+L
D
∆
˙
i
D
+L
AD
∆
˙
i
F
= 0 , (5.61)
in which the expression for ∆
˙
i
F
in (5.60) is inserted giving
r
D
∆i
D
+L
D
∆
˙
i
D
+L
AD
∆u
F
−r
F
∆i
F
−L
AD
∆
˙
i
D
L
F
= 0 . (5.62)
This equation can be rewritten as
∆
˙
i
D
+
r
D
L
D
−L
2
AD
/L
F
∆i
D
= −∆u
F
L
AD
/L
F
L
D
−L
2
AD
/L
F
+r
F
L
AD
L
F
∆i
F
. (5.63)
As pointed out above the current in the damper circuit will decay must
faster, which means that the current in the ﬁeld winding, ∆i
F
, can be re
garded as a constant in Equation (5.63). The time constant of the decay of
the damper winding current, ∆i
D
, is the subtransient time constant of the
open circuit in the d–axis, T
′′
do
, and can now with the above assumptions be
identiﬁed from Equation (5.63) as
T
′′
do
=
L
D
−L
2
AD
/L
F
r
D
. (5.64)
When i
D
has vanished, i.e. when i
D
= 0, the ﬁeld current is determined
solely by the upper equation in (5.57), and the transient time constant of
the open circuit, T
′
do
, and together with the lower equation of (5.58) it is
easily seen that this time constant is given by
T
′
do
= L
F
/r
F
. (5.65)
Accordingly, for a synchronous machine with a damper winding in the q–
axis,
T
′′
qo
= L
Q
/r
Q
(5.66)
can be derived.
76 5. Synchronous Machine Model
Round Rotor Salient Pole
x
d
(p.u.) 1.0 – 2.3 0.6 – 1.5
x
q
(p.u.) 1.0 – 2.3 0.4 – 1.0
x
′
d
(p.u.) 0.15 – 0.4 0.2 – 0.5
x
′′
d
(p.u.) 0.12 – 0.25 0.15 – 0.35
x
′
q
(p.u.) 0.3 – 1.0 —
x
′′
q
(p.u.) 0.12 – 0.25 0.2 – 0.45
T
′
do
(s) 3.0 – 10.0 1.5 – 9.0
T
′′
do
(s) 0.02 – 0.05 0.01 – 0.05
T
′
qo
(s) 0.5 – 2.0 –
T
′′
qo
(s) 0.02 – 0.05 0.01 – 0.09
H (s) 3 – 5 (n = 3000 rpm) 1.5 – 5
5 – 8 (n = 1500 rpm)
Table 5.1. Typical values of some parameters of synchronous machines.
The quantities introduced in this and the previous sections are important
parameters of a synchronous machine and are usually given by the manufac
turer of the machine. The reasons for this are that they are easily measured
by simple tests and that they are introduced in a natural way into the sim
pliﬁed models we will derive in the next section. In Table 5.1, typical values
of the discussed parameters for diﬀerent types and sizes of synchronous ma
chines are given. Reactances instead of inductances are given in Table 5.1,
i.e. x
d
= ω
0
L
d
etc., with ω
0
= 2πf
0
. The p.u. base is the power and voltage
ratings of the machine.
5.6 Simpliﬁed Models of the Synchronous Machine
5.6.1 Derivation of the fourthorder model
A complete and exact model of the synchronous machine, considering the as
sumptions and approximations made, was presented in sections 5.2 and 5.3.
When a synchronous machine is modelled in for example a software package
for stability simulations, these are often the equations used to represent the
synchronous machine. Nevertheless, it is, for several reasons, often mean
ingful to use models that comprise more simpliﬁcations and approximations
than those derived earlier. If a better insight into a problem is wanted or if
a problem is analyzed without using computer simulations, simpliﬁed repre
sentations of the synchronous machine are often the only possibility. Here,
such a model will be derived under the following assumptions:
• The stator transients are neglected, i.e. we set
˙
Ψ
d
=
˙
Ψ
q
= 0.
5.6. Simpliﬁed Models of the Synchronous Machine 77
• The d–axis contains no damper windings.
• In the q–axis, one damper winding is modelled.
With these assumptions, the synchronous machine is described by the
equations
_
¸
¸
_
¸
¸
_
Ψ
d
= L
d
i
d
+L
AD
i
F
,
Ψ
q
= L
q
i
q
+L
AQ
i
Q
,
Ψ
F
= L
F
i
F
+L
AD
i
d
,
Ψ
Q
= L
Q
i
Q
+L
AQ
i
q
,
(5.67)
_
u
F
= r
F
i
F
+
˙
Ψ
F
,
0 = r
Q
i
Q
+
˙
Ψ
Q
,
(5.68)
_
u
d
= −ri
d
+ωΨ
q
,
u
q
= −ri
q
−ωΨ
d
.
(5.69)
For completeness, some relations deﬁned above are repeated here.
_
L
′
d
= L
d
−L
2
AD
/L
F
L
′′
q
= L
q
−L
2
AQ
/L
Q
(5.70)
The goal is now to eliminate all quantities with indices F and Q, except for
u
F
, from the equations above to get a model that can be used to represent
the synchronous machine as a component in a system. That means that the
only quantities that should be present in the model are stator voltages, the
stator current, and u
F
, which is a control variable that can be changed by
the excitation system as described in Chapter 6.
From the Equations (5.67), i
Q
is eliminated, which leads to
Ψ
q
−
L
AQ
L
Q
Ψ
Q
= L
′′
q
i
q
. (5.71)
Now, by deﬁning
e
′
d
= ω
L
AQ
L
Q
Ψ
Q
, (5.72)
Equation (5.71) can be written as
Ψ
q
−e
′
d
/ω = L
′′
q
i
q
. (5.73)
Equivalent derivations can be done for the rotor circuit in the d–axis,
i.e. for the ﬁeld winding.
Ψ
d
−
L
AD
L
F
Ψ
F
= L
′
d
i
d
(5.74)
is obtained, which, using the deﬁnition
e
′
q
= −ω
L
AD
L
F
Ψ
F
, (5.75)
78 5. Synchronous Machine Model
can be written as
Ψ
d
+e
′
q
/ω = L
′
d
i
d
. (5.76)
Now, the quantities
_
e
d
= ωL
AQ
i
Q
,
e
q
= −ωL
AD
i
F
(5.77)
are introduced. From the equations above, the relationships
_
e
d
= e
′
d
−(x
q
−x
′′
q
)i
q
,
e
q
= e
′
q
+ (x
d
−x
′
d
)i
d
,
(5.78)
with x
d
= ωL
d
etc. can be obtained.
According to Equations (5.77) and (5.78),
i
Q
=
e
′
d
−(x
q
−x
′′
q
)i
q
ωL
AQ
. (5.79)
Substituting this into the second equation of (5.68), which together with
(5.72) give
r
Q
e
′
d
−(x
q
−x
′′
q
)i
q
ωL
AQ
+
L
Q
ωL
AQ
˙ e
′
d
= 0 , (5.80)
where it has been assumed that ˙ ω = 0 or ω = ω
0
. Equation (5.80) can also
be written as
T
′′
qo
˙ e
′
d
+e
′
d
−(x
q
−x
′′
q
)i
q
= 0 . (5.81)
For the rotor circuit in the d–axis, i.e. for the exciter winding, accordingly
T
′
do
˙ e
′
q
+e
′
q
+ (x
d
−x
′
d
)i
d
= −
ωL
AD
r
F
u
F
(5.82)
is obtained, which, using
e
F
=
ωL
AD
r
F
u
F
, (5.83)
can be written as
T
′
do
˙ e
′
q
+e
′
q
+ (x
d
−x
′
d
)i
d
= −e
F
. (5.84)
If Equations (5.81) and (5.84) are Laplace–transformed and the voltage
equation (5.69) is rewritten with the introduced quantities, we obtain
_
∆u
d
∆u
q
_
=
_
∆e
′
d
∆e
′
q
_
+
_
−r
d
x
′′
q
−x
′
d
−r
q
__
∆i
d
∆i
q
_
, (5.85)
with e
′
d
given by
∆e
′
d
=
(x
q
−x
′′
q
)∆i
q
1 +sT
′′
qo
(5.86)
5.6. Simpliﬁed Models of the Synchronous Machine 79
and e
′
q
from
∆e
′
q
= −
∆e
F
+ (x
d
−x
′
d
)∆i
d
1 +sT
′
do
. (5.87)
This model, i.e. Equations (5.85)–(5.87), together with the swing equation,
are often called the fourth order model. This model of the synchronous
machine demands, with the assumptions made here, four state variables:
∆e
′
d
and ∆e
′
q
and the “mechanical” quantities ∆ω and ∆δ.
Often, also the damper winding in the q–axis can be neglected. That
leads to a third order model, which can be written as
_
∆u
d
∆u
q
_
=
_
0
∆e
′
q
_
+
_
−r
d
x
q
−x
′
d
−r
q
__
∆i
d
∆i
q
_
, (5.88)
with ∆e
′
q
according to Equation (5.87).
5.6.2 The HeﬀronPhillips formulation for stability studies
The thirdorder model of the synchronous machine derived in the previous
section can be formulated as a block diagram. The basis for the model
presented here, which was originally proposed by Heﬀron and Phillips, is
the ”single machine, inﬁnite bus” (SMIB) setup. By introducing a number
of new constants, a very compact notation is achieved. The model shall
enable the reader to directly implement a usable representation of an SMIB
system, including the mechanical dynamics, ﬁeld winding, and excitation
system. This implementation can be used directly for stability studies. In
the present case, a generic simpliﬁed representation of the excitation system
is used. Detailed descriptions and common variants of these systems can be
found in chapter 6.
Background: Lowfrequency oscillations in power systems
In large interconnected power systems, lowfrequency oscillations can arise
spontaneously. Unlike the electromechanical oscillations described by the
swing equation, these oscillations have a frequency of well below 1 Hz (about
3 – 10 cycles per minute depending on the power system). They are known
to either decay slowly or continue to grow, which leads eventually to system
separation. The main cause of the oscillations is a lack of damping in the
system due to the eﬀect of excitation systems in synchronous machines.
An eﬀective countermeasure to lowfrequency oscillations is the usage of
Power Systems Stabilizers (PSS). These are automatic controllers acting on
the excitation systems, providing an additional damping component which
leads to an attenuation of the oscillations. Further details about this system,
which can be directly attached to the HeﬀronPhillips model derived in this
section, will be presented in the chapters 6 and 7.
80 5. Synchronous Machine Model
Block diagram
Figure 5.5 presents the transfer function block diagram of the Heﬀron
Phillips model. Note that all quantities are presented in per unit. The
mechanical system is represented by the system inertia and the damping
constant, where the torque balance ∆T
m
− ∆T
e
is considered as an input
and the incremental torque angle ∆δ as an output. The electrical part of
the system consists of three main parts: 1) the composition of the electri
cal torque (inﬂuenced by ∆δ over constant K
1
and the internal incremental
voltage ∆e
′
q
over constant K
2
), 2) the eﬀect of the ﬁeld winding (determined
by the ﬁeld winding constant K
3
and inﬂuenced by ∆δ over constant K
4
),
and 3) the eﬀect of the excitation system (inﬂuenced by ∆δ over constant
K
5
and ∆e
′
q
over constant K
6
). The excitation system itself is modeled by
a ﬁrstorder transfer function including the ampliﬁcation factor K
A
and the
time constant T
A
.
5.6. Simpliﬁed Models of the Synchronous Machine 81
G
'
m
T
'
e
T
'
Z
'
t
u
'
' q
e
'
F
e
'
0
s
e
t
t
u
'
M
e
c
h
a
n
i
c
a
l
t
o
r
q
u
e
c
h
a
n
g
e
S
y
s
t
e
m
i
n
e
r
t
i
a
1
2
H
s
0
2
f
s
S
D
K
D
a
m
p
i
n
g
c
o
n
s
t
a
n
t
1
K
2
K
5
K
1
A
A
K
T
s
6
K
4
K
3
'
3
1
d
o
K
T
K
s
I
n
f
l
u
e
n
c
e
o
f
t
o
r
q
u
e
a
n
g
l
e
o
n
t
e
r
m
i
n
a
l
v
o
l
t
a
g
e
I
n
f
l
u
e
n
c
e
o
f
i
n
t
e
r
n
a
l
v
o
l
t
a
g
e
o
n
t
e
r
m
i
n
a
l
v
o
l
t
a
g
e
F
i
e
l
d
w
i
n
d
i
n
g
E
x
c
i
t
a
t
i
o
n
s
y
s
t
e
m
S
y
n
c
h
r
o
n
o
u
s
s
p
e
e
d
I
n
f
l
u
e
n
c
e
o
f
t
o
r
q
u
e
a
n
g
l
e
o
n
e
l
e
c
t
r
i
c
t
o
r
q
u
e
I
n
f
l
u
e
n
c
e
o
f
i
n
t
e
r
n
a
l
v
o
l
t
a
g
e
o
n
e
l
e
c
t
r
i
c
t
o
r
q
u
e
I
n
f
l
u
e
n
c
e
o
f
t
o
r
q
u
e
a
n
g
l
e
o
n
f
i
e
l
d
v
o
l
t
a
g
e
V
o
l
t
a
g
e
s
e
t
p
o
i
n
t
c
h
a
n
g
e
F
i
g
u
r
e
5
.
5
.
B
l
o
c
k
d
i
a
g
r
a
m
r
e
p
r
e
s
e
n
t
i
n
g
s
y
n
c
h
o
n
o
u
s
m
a
c
h
i
n
e
i
n
a
S
M
I
B
s
y
s
t
e
m
.
82 5. Synchronous Machine Model
Derivation of constants K
1
. . . K
6
The constants K
1
. . . K
6
shown in Figure 5.5 describe internal inﬂuence fac
tors within the system and can be found by a comparison of coeﬃcients with
the equations governing the synchronous machine dynamics. While K
1
and
K
2
are derived from the computation of the electric torque, K
3
and K
4
have
their origin in the ﬁeld voltage equation (5.87). K
5
and K
6
come from the
equation governing the terminal voltage magnitude. Below we will calculate
these constants from the previously presented thirdorder model.
SMIB system setup: The power system under consideration is a socalled
”single machine, inﬁnite bus” (SMIB) system. This means that one syn
chronous generator is connected via a transmission line (represented by series
impedance Z = R+jX and shunt admittance Y = G+jB) to an ”inﬁnite”
bus with constant voltage u
0
. The generator has a terminal voltage u
t
and
injects the armature current i via its terminals. For the calculations below,
the armature currents i
d
and i
q
are assumed to be known.
The current and voltage phasors are deﬁned as shown in Figure 5.6. The
torque angle δ is deﬁned as the angle between u
0
and the internal voltage
e
′
q
. The following deﬁnitions will be used later in the computations:
i = i
d
+ji
q
, v
t
= v
d
+jv
q
,
1 +ZY = C
1
+jC
2
,
R
1
= R −C
2
x
′
d
, X
1
= X +C
1
x
q
,
X
2
= X +C
1
x
′
d
, R
2
= R −C
2
x
q
,
Z
2
e
= R
1
R
2
+X
1
X
2
,
Y
d
= (C
1
X
1
−C
2
R
2
)/Z
2
e
, Y
q
= (C
1
R
1
+C
2
X
2
)/Z
2
e
,
(5.89)
The voltages and current in the SMIB system are related to the admittances
and impedances by
i = Y u
t
+Z
−1
(u
t
−u
0
) or Zi = (1 +ZY )u
t
−u
0
. (5.90)
This can be separated into real and imaginary parts, which yields
_
R −X
X R
_ _
i
d
i
q
_
=
_
C
1
−C
2
C
2
C
1
_ _
u
d
u
q
_
−u
0
_
sin δ
cos δ
_
, (5.91)
where
C
1
= 1 +RG−XB , C
2
= XG+RB . (5.92)
Now, Equation (5.88) is taken into account with one further simpliﬁcation:
the armature resistance is neglected, i.e. r
d
= r
q
= 0. This can be inserted
into Equation (5.91), which yields
_
i
d
i
q
_
=
_
Y
d
Y
q
_
e
′
q
−
u
0
Z
2
e
_
R
2
X
1
−X
2
R
1
_ _
sinδ
cos δ
_
. (5.93)
5.6. Simpliﬁed Models of the Synchronous Machine 83
qaxis
e‘
q
u
t
i
į u
0
daxis
Figure 5.6. Phasor diagram of SMIB system.
This can be linearized again, resulting in
_
∆i
d
∆i
q
_
=
_
Y
d
Y
q
_
∆e
′
q
+
_
F
d
F
q
_
∆δ , (5.94)
with the constants F
d
and F
q
introduced for abbreviation, which are equal
to
_
F
d
F
q
_
=
u
0
Z
2
e
_
−R
2
X
1
X
2
R
1
_ _
cos δ
0
sinδ
0
_
, (5.95)
where δ
0
is the initial torque angle.
Now, the constants K
1
. . . K
6
can be deﬁned in terms of the introduced
notation.
K
1
and K
2
from electric torque: The electric torque of a synchronous
machine can be approximated by the electric power for small deviations
from the synchronous speed. The overall power is the sum of powers in the
d and qaxis:
T
e
≈ P
e
= i
d
u
d
+i
q
u
q
. (5.96)
Taking into account Equation (5.88) with r
d
= r
q
= 0, Equation (5.96) can
be written as
T
e
= i
q
e
′
q
+ (x
q
−x
′
d
)i
d
i
q
. (5.97)
Merging this with Equation (5.94), this yields with one further linearization:
∆T
e
= K
1
∆δ +K
2
∆e
′
q
, (5.98)
where the wanted constants K
1
and K
2
are equal to:
_
K
1
K
2
_
=
_
0
i
q0
_
+
_
F
d
F
q
Y
d
Y
q
_ _
(x
q
−x
′
d
)i
q0
e
′
q0
+ (x
q
−x
′
d
)i
d0
_
. (5.99)
84 5. Synchronous Machine Model
K
3
and K
4
from ﬁeld voltage equation: The ﬁeld winding circuit voltage
equation from (5.84) can be linearized, which yields
(1 +s T
′
do
)∆e
′
q
= −∆e
F
−(x
d
−x
′
d
)∆i
d
. (5.100)
Substituting ∆i
d
of Equation (5.94) into this yields
(1 +s T
′
do
K
3
)∆e
′
q
= −K
3
(∆e
F
+K
4
∆δ) (5.101)
where
K
3
= 1/(1 + (x
d
−x
′
d
)Y
d
) (5.102)
K
4
= (x
d
−x
′
d
)F
d
. (5.103)
K
5
and K
6
from terminal voltage magnitude: The magnitude of the ter
minal voltage of the generator can be expressed by the d and q components:
u
2
t
= u
2
d
+u
2
q
. (5.104)
The deviation from a steady state is thus equal to
∆u
t
= (u
d0
/u
t0
)∆u
d
+ (u
q0
/u
t0
)∆u
q
. (5.105)
Substituting Equation (5.94) into Equation (5.105) yields
∆u
t
= K
5
∆δ +K
6
∆e
′
q
, (5.106)
where
_
K
5
K
6
_
=
_
0
u
q0
/u
t0
_
+
_
F
d
F
q
Y
d
Y
q
_ _
−x
′
d
u
q0
/u
t0
x
q
u
d0
/u
t0
_
. (5.107)
Note that all subscripts 0 denote steadystate quantities in this section.
Excitation system parameters: The only missing parameters for the com
pletion of the HeﬀronPhillips model are the excitation system parameters
K
A
and T
A
. The corresponding transfer function represents a generic exciter
and voltage regulator of the fastresponse type. Reasonable values for the
parameters are K
A
≈ 50 and T
A
≈ 0.05 s.
6
Voltage Control in Power Systems
Having introduced the principal governing equations for the synchronous ma
chine in the previous chapter, this chapter deals with the basics of voltage
control in electric power systems. First, the relation between voltage and
the reactive power balance is discussed and various inﬂuences to this balance
are presented. The primary voltage control equipment in the synchronous
machine is also discussed and ﬁnally a short description of the hierarchical
voltage control structure is provided. The following structure is based on
[6]
6.1 Relation between voltage and reactive power
As explained in detail in chapters 2 – 4, the active power balance of a power
system must be maintained in order to keep the system in steady state.
Furthermore, the reactive power balance must be kept in such a way that
the voltages are within the acceptable limits. If the active power balance is
not kept, the frequency in the system will be inﬂuenced, while an improper
reactive power balance will result in deviations of the voltages in the system
from the desired ones.
1
Normally the power system is operated such that the voltage drops along
the lines are small. The node voltages of the system will then almost be equal
(ﬂat voltage proﬁle). In this case the transmission system is eﬀectively used,
i.e. primarily for transmission of active power, and not for transmission of
reactive power.
Thus, the voltage magnitudes can be controlled to desired values by
control of the reactive power. Increased production of reactive power results
in higher voltage near the production source, while an increased consumption
of reactive power results in lower voltage. While the active power is entirely
produced by the generators in the system, there are several sources and sinks
of reactive power. On the other hand, the reactive power, in contrast with
the active power, cannot be transported over long distances in the system,
1
The frequency deviation is a consequence of an imbalance between power fed into the
system by the prime movers and the electric power consumed by the loads and losses.
However, the generated reactive power is always equal to the consumed reactive power,
which is a consequence of Kirchhoﬀ’s laws. The voltage magnitudes are always adjusted
such that this balance is maintained. If the voltages settle at too low values, the reason
is that the reactive generation is too small. Vice versa, overvoltages arise when reactive
generation is too high. However, this does not mean that the generated and consumed
reactive power is not equal as in the case of active power.
85
86 6. Voltage Control in Power Systems
since normally X >> R in a power system. Consequently, the reactive
power can be regarded as a fairly local quantity.
Important generators of reactive power are:
• Overexcited synchronous machines
• Capacitor banks
• The capacitance of overhead lines and cables
• FACTS devices
2
Important consumers of reactive power are:
• Inductive static loads
• Underexcited synchronous machines
• Induction motors
• Shunt reactors
• The inductance of overhead lines and cables
• Transformer inductances
• FACTS devices
For some of these, the reactive power is easy to control, while for others
it is practically impossible. The reactive power of the synchronous machines
is easily controlled by means of the excitation. Switching of shunt capacitors
and reactors can also control the reactive power. FACTS devices oﬀer also
a possibility to control the reactive power.
It is most eﬀective to compensate the reactive power as close as possible
to the reactive load. There are certain high voltage tariﬀs to encourage large
consumers, e.g. industries, and electrical distributions companies to compen
sate their loads in an eﬀective way. These tariﬀs are generally designed so
that the reactive power is only allowed to reach a certain percentage of the
active power. If this percentage is exceeded, the consumer has to pay for
the reactive power. The high voltage network is in this way primarily used
for transmission of active power.
The reactive losses of power lines and transformers depend on the size of
the reactance. For overheadtransmission lines the reactance can be slightly
reduced by the use of multiple conductors. The only possibility to radi
cally reduce the total reactance of a transmission line s to connect a series
capacitor or a series FACTS device.
2
FACTS = Flexible AC Transmission Systems. FACTS devices are power electronics
devices equipped with fast control that can often can be used for reactive power control.
They can normally be used both as reactive power sources and sinks. Further discussion
of FACTS devices is given in the Section 6.3.4.
6.2. Voltage Control Mechanisms 87
6.2 Voltage Control Mechanisms
The following factors inﬂuence primarily the voltages in a power system:
• Terminal voltages of synchronous machines
• Impedances of lines
• Transmitted reactive and active power
• Turns ratio of transformers
A suitable use of these leads to the desired voltage proﬁle.
The generators are often operated at constant voltage, by using an auto
matic voltage regulator (AVR). The output from this controls the excitation
of the machine via the electric ﬁeld exciter. In that way, the voltage can
be kept equal to the set value. The voltage drop caused by the generator
transformer is sometimes compensated totally or partly by this means, and
the voltage can consequently be kept constant on the high voltage side of
the transformer. Synchronous compensators can also be installed for volt
age control. These are synchronous machines without turbine or mechanical
load, which can produce and consume reactive power by controlling the ex
citation. Nowadays new installations of synchronous compensators are very
rare, and power electronics based solutions are preferred if fast voltage con
trol is needed.
The reactive power transmitted over a line has a great impact on the
voltage proﬁle. Large reactive transmissions cause large voltage drops, thus
these should be avoided. Instead, the production of reactive power should
be as close as possible to the reactive loads. This can be made by the exci
tation of the synchronous machines, as described above. However, there are
often no synchronous machines close to the load, so the most costeﬀective
way is to use shunt capacitors which are switched according to the load vari
ations. A power electronics based device can be economically motivated if
fast response or accuracy in the regulation is required. Shunt reactors must
sometimes be installed to limit the voltages to reasonable levels. In net
works which contain a lot of cables this is also necessary, since the reactive
generation from these is much larger than from overhead lines. (C is much
larger and X is smaller.)
88 6. Voltage Control in Power Systems
6.3 Primary Voltage Control
The task of the primary voltage control is to control the reactive output
from a device so that the voltage magnitude is kept at or close to the set
value of the controller. Usually the node of the controlled voltage is at the
same or very close to the node of the reactive device, since reactive power is
a fairly local quantity. The set values for the voltage controllers are selected
so that the desired voltage proﬁle of the system is obtained. The selection
of set values is the task of the secondary voltage control, which is brieﬂy
summarized in section 6.4.
In the following the most important devices for reactive power and volt
age control are described.
6.3.1 Synchronous Machine Excitation System and AVR
The reactive power output of synchronous machine can for a given active
power level be adjusted within the limits of the capability curve by the
excitation system. This oﬀers a very eﬃcient and fast way to control the
terminal voltage of the machine, which in most power system is the most
important voltage control.
The main purpose of the excitation system is to feed the ﬁeld winding
of the synchronous machine with direct current so that the main ﬂux in the
rotor is generated. Further, the terminal voltage of the synchronous ma
chine is controlled by the excitation system, which also performs a number
of protection and control tasks. A schematic picture of a generator with
excitation system is depicted in Figure 6.1. Below, a short description of
the functions of the diﬀerent blocks in Figure 6.1 is given:
• The exciter supplies the ﬁeld winding with direct current and thus
comprises the “power part” of the excitation system.
• The controller treats and ampliﬁes the input signals to a level and
form that is suited for the control of the exciter. Input signals are pure
control signals as well as functions for stabilizing the exciter system.
• The voltage measurement and load compensation unit mea
sures the terminal voltage of the generator and rectiﬁes and ﬁlters it.
Further, load compensation can be implemented if the voltage in a
point apart from the generator terminals, such as in a ﬁctional point
inside the generator’s transformer, should be kept constant.
• The power system stabilizer, PSS, gives a signal that increases
the damping to the controller, cf. Chapter 7. Usual input signals for
the PSS are deviations in rotor speed, accelerating power, or voltage
frequency.
6.3. Primary Voltage Control 89
Synchronous
Machine
PSS
Measurements
Voltage
Load Compensation
Limiter and
Protection
Exciter Regulator
A
A
= To the Power System
Figure 6.1. Schematic picture of a synchronous machine with excita
tion system with several control, protection, and supervisory functions.
• The limiter and protection can contain a large number of functions
that ensure that diﬀerent physical and thermal limits, which genera
tor and exciter have, are not exceeded. Usual functions are current
limiters, over–excitation protection, and under–excitation protection.
Many of these ensure that the synchronous machine does not produce
or absorb reactive power outside of the limits it is designed for.
Today, a large number of diﬀerent types of exciter systems is used. Three
main types can be distinguished:
• DC excitation system, where the exciter is a DC generator, often
on the same axis as the rotor of the synchronous machine.
• AC excitation system, where the exciter is an AC machine with
rectiﬁer.
• Static excitation system, where the exciting current is fed from a
controlled rectiﬁer that gets its power either directly from the gen
erator terminals or from the power plant’s auxiliary power system,
normally containing batteries. In the latter case, the synchronous ma
chine can be started against an unenergised net, “black start”. The
batteries are usually charged from the net.
Below, a more comprehensive treatment of some of the functions de
scribed above is given.
90 6. Voltage Control in Power Systems
U
T
I
T
U
c
U
T
R
c
j X
c
+ ( )I
T
+ =
U
c
1
1 sT
R
+


+
U
err
Σ
U
ref
Figure 6.2. Block diagram of compensating circuit.
Load Compensation Equipment
Figure 6.2 shows the block diagram of a compensation circuit, consisting of
a converter for measured values, a ﬁlter, and a comparator.
There are several reasons for the use of compensation in voltage control
of synchronous machines. If two or more generators are connected to the
same bus, the compensation equipment can be used to create an artiﬁcial
impedance between those. That is necessary to distribute the reactive power
in an appropriate way between the machines. The voltage is measured
“somewhat inside” the generator, corresponding to positive values of R
c
and X
c
in Figure 6.2. If a machine is connected with a comparatively large
impedance to the system, which usually is the case since the generator’s
transformer normally has an impedance in the order of magnitude of 10%
on basis of the machine, it can be desirable to compensate a part of this
impedance by controlling the voltage “somewhat inside” of that impedance.
This then corresponds to negative values of R
c
and X
c
. As a rule, X
c
is
much larger than R
c
.
DC Excitation Systems
Today, hardly any DC excitation systems are being installed, but many of
these systems are still in operation. Generally, it can be said that there is
a large number of variants of the diﬀerent excitation systems listed above.
Every manufacturer uses its own design, and demands that depend on the
application often lead to considerable diﬀerences in the detailed models of
the devices in each group. Here, typical examples for models will be given. In
reality, the models given by the manufacturers and power suppliers must be
used. One example of a DC excitation system, the IEEE type DC1 system,
is given in Figure 6.3. The input signal for the controller is the voltage
error U
err
from the compensation equipment. The stabilizing feedback U
F
is subtracted, and sometimes a signal from the PSS is added. Both these
signals vanish in steady state. The controller is mainly described by the
dominating time constant T
A
and the ampliﬁcation K
A
. The limits can
represent saturation eﬀects or limitations of the power supply. The time
6.3. Primary Voltage Control 91
1 sT
C
+
1 sT
D
+

K
A
1 sT
A
+

1
sT
E

S
E
E
F
( ) K
E
+
sK
F
1 sT
F
+

U
RMAX
U
RMIN
Σ
Σ
U
R
U
FE
+

+
+

E
F
U
F
U
err
U
PSS
Figure 6.3. Model of DC exciter system (IEEE Type DC1).
constants T
C
and T
D
can be used to model internal time constants in the
controller. These are often small and can then usually be neglected.
The output signal from the voltage controller, U
R
, controls the exciter.
The exciter consists of a DC machine that can be excited independently or
shunt excited. For shunt excited machines, the parameter K
E
models the
setting of the ﬁeld regulator. The term S
E
represents the saturation of the
exciter and is a function of the exciter’s output voltage, E
F
. If saturation is
neglected, that is S
E
= 0, the eﬀective time constant of the exciter becomes
T
E
/K
E
, and its eﬀective ampliﬁcation is 1/K
E
.
AC Excitation Systems
For AC excitation systems, the exciter consists of a smaller synchronous ma
chine that feeds the exciter winding through a rectiﬁer. The output voltage
of the exciter is in this case inﬂuenced by the loading. To represent these
eﬀects, the exciter current is used as an input signal in the model. In Fig
ure 6.4, an example of a model of AC exciter systems is shown (IEEE type
AC1). The structure of the model is basically the same as for the DC exci
tation system. Some functions have been added. The rectiﬁer of the exciter
prevents (for most exciters) the exciter current from being negative. The
feedback with the constant K
D
represents the reduction of the ﬂux caused
by a rising ﬁeld current I
F
. That constant depends on the synchronous and
transient reactances of the exciter. The voltage drop inside the rectiﬁer is
described by the constant K
C
, and its characteristic is described by F
EX
,
which is a function of the load current.
DC and AC excitation systems are sometimes called rotating exciters,
since they contain rotating machines. That distinguishes them from static
excitation systems, which are described in the sequel.
92 6. Voltage Control in Power Systems
1 sT
C
+
1 sT
D
+

K
A
1 sT
A
+

1
sT
E

F
EX
f I
H
( ) =
I
H
K
C
I
F
U
E
 =
K
E
S
E
+
K
D
sK
F
1 sT
F
+

E
F
F
EX
U
FE
U
F
U
R
U
RMAX
U
RMIN
U
PSS
U
err
Σ
Σ
Σ
U
E
I
F
+
+
+
+
+

Π

0
Figure 6.4. Model of an AC exciter system (IEEE Type AC1).
Static Excitation Systems
In static excitation systems, the exciter winding is fed through a transformer
and a controlled rectiﬁer. By far most exciter systems installed today are of
that type, and a large number of variants exists. The primary voltage source
can be a voltage transformer that is connected to the generator terminals,
but even a combination of voltage and current transformers can be found.
With the latter arrangement, an exciter current can be obtained even if the
voltage at the generator terminals is low, for example during a ground fault
in or near the power plant. Sometimes, it is possible to supplement these
voltage sources by using the auxiliary power of the power plant as voltage
source. That makes it possible to start the generator in an unenergised net.
An example of a model of a static exciter system is shown in Figure 6.5.
Static excitation systems can often deliver negative ﬁeld voltage and
even negative ﬁeld current. However, the maximum negative ﬁeld current is
usually considerably lower than the maximum positive ﬁeld current.
The time constants are often so small that a stabilizing feedback is not
needed. The constant K
F
can then be set to zero. Since the exciter system
is normally supplied directly from the generator bus, the maximum exciter
voltage depends on the generator’s output voltage (and possibly its current).
This is modelled by the dependency of the limitations of the exciter output
on the generator’s output voltage. The constant K
C
represents the relative
voltage drop in the rectiﬁer.
6.3. Primary Voltage Control 93
1 sT
C
+
1 sT
D
+

K
A
1 sT
A
+

sK
F
1 sT
F
+

E
F
U
T
U
RMAX
K
C
I
FE

U
T
U
RMIN
K
C
I
FE

U
IMAX
U
IMIN
U
err
U
PSS
Σ
+
+

U
F
Figure 6.5. Model of a static exciter system.
6.3.2 Reactive Shunt Devices
In many systems, the reactive control capabilities of the synchronous ma
chines are not suﬃcient to keep the voltage magnitudes within prescribed
limits at all loading conditions. The varied loading condition of the system
during low and peak load situations implies that the reactive power needed
to keep the desired voltage magnitudes vary signiﬁcantly. It is impossible
and also unwise to use the reactive power capabilities of synchronous ma
chines to compensate for this, since many synchronous machines will be
driven to their capability limits, and the fast and continuous reactive power
control oﬀered by the synchronous machine will not be available. Therefore,
in most systems, breakerswitched shunt capacitor banks and shunt reac
tors are used for a coarse control of reactive power, so that the synchronous
generators can be used for the fast and continuous control.
Generally shunt capacitors are switched on during highload conditions,
when the reactive consumption from loads and line reactances is the highest.
Since these load variations are rather slow and predictable, no fast control
is needed and the capacitor banks can be breaker switched. In systems with
long highvoltage lines, the reactive power generation of these lines can be
very high during light load conditions. To keep the voltages at acceptable
levels, shunt reactors might be needed. The sizes of the reactive shunt el
ements determine how accurate the control can be. Capacitor banks, in
particular, can often be switched in smaller units, while shunt reactors are
most often installed in one single unit, because of high costs. A factor limit
ing the element size is the transient voltage change resulting from switching.
The fundamental frequency voltage change in pu caused by switching a shunt
element can be estimated as ∆V = Q
shunt
/S
sc
, where Q
shunt
is the size of
the shunt element and S
sc
is the shortcircuit power at the node.
Reactive shunt elements are also used for reactive power and voltage con
trol at HVDC terminal stations. Since typically a line commutated HVDC
converter station consumes about 50% as much reactive power as active
94 6. Voltage Control in Power Systems
power transmitted, reactive compensation is needed. Part of the reactive
compensation is usually provided by harmonic ﬁlters needed to limit the
harmonic current injection into the ac networks. These ﬁlters are almost
purely capacitive at fundamental frequency.
6.3.3 Transformer Tap Changer Control
An important method for controlling the voltage in power systems is by
changing the turns ratio of transformers. Certain transformers are equipped
with a number of taps on one of the windings. Voltage control can be ob
tained by switching between these taps, as illustrated in Figure 6.6. Switch
ing during operation by means of tap changers is very eﬀective and useful for
voltage control. Normally, the taps are placed on the highvoltage winding
(the upper side), since then lower currents needs to be switched.
N
2
N
1
U
1
Z
k
τU
2
U
2
Transformer with variable turns ratio (tap changer).
Figure 6.6. Transformer with a variable turns ratio (tap changer).
If N
1
is the number of turns on the highvoltage side and N
2
is the
number of turns on the lowvoltage side, the turns ratio of transformer is
deﬁned as
τ =
N
1
N
2
. (6.1)
Then, the relation between the voltage phasors on the highvoltage side U
1
and on the lowvoltage side U
2
, at no load, is
U
2
=
U
1
τ
. (6.2)
If the voltage decreases on the highvoltage side, the voltage on the lower side
can be kept constant by decreasing, that is, by switching out a number of
windings on the highvoltage side. When the transformer is loaded, equation
(6.2) is of course incorrect, since the load current yields a voltage drop over
the leakage reactance of the transformer Z
k
, but the same principle can still
be applied for voltage control.
Transformers with automatic tap changer control are often used for volt
age control in distribution networks. The voltage at the consumer side can
therefore be kept fairly constant even though voltage variations occur on the
highvoltage network. Time constants in these regulators are typically in the
6.3. Primary Voltage Control 95
order of tenths of seconds. In some transformers, the turns ratio cannot be
changed during operation, but just manually when the transformer is un
loaded. In this case, voltage variations in the network cannot be controlled,
with voltage levels changing in large steps.
6.3.4 FACTS Controllers
As previously mentioned, power electronicsbased equipment, usually re
ferred to as FACTS controllers or devices, can be used for fast voltage and
reactive power control. The ﬁrst devices introduced were based on thyristors
as active elements, but more recently, devices using voltage source converters
have been introduced; with the latter, a faster and more powerful control
can be achieved. FACTS devices used for voltage control are connected
in shunt. Two such devices are discussed here: the Static var compensator
(SVC), and the STATic synchronous COMpensator (STATCOM). The SVC
is based on thyristors, while the STATCOM is based on a voltage source con
verter. There are also seriesconnected FACTS devices that in principle can
be used for voltage control, but the main reason for installing these devices
is to control active power, which indirectly inﬂuences voltages. Among these
devices, the uniﬁed power ﬂow controller (UPFC) is equipped with a shunt
part for voltage control, which basically operates as a STATCOM.
i
s
V
Thyristor controlled reactor.
Figure 6.7. Thyristor controlled reactor.
96 6. Voltage Control in Power Systems
Static Var Compensator
The SVC may be composed of two diﬀerent shunt elements, that is, a thyris
tor controlled reactor (TCR) and thyristor switched capacitor (TSC) banks.
If fast switching of the capacitor banks is not needed, one can also use
breakerswitched capacitors. The TCR is depicted in Figure 6.7 by delaying
the ﬁring of the thyristors, a continuous control of the current through the
reactor can be obtained, with the reactive power consumption varying be
tween 0 and V
2
/X, where X is the reactance of the reactor. By combining
the TCR with a suitable number of capacitor banks, a continuous control
of the reactive power can be achieved by a combination of capacitor bank
switching and control of the reactor current. Usually, the TCR and TSC
are connected to the highvoltage grid through a transformer, as shown in
Figure 6.8. The control system of the SVC controls the reactive output so
that the voltage magnitude of the controlled node is kept constant. Usually,
a certain slope is introduced in the control, as shown in Figure 6.9, which
shows the reactive current as a function of the voltage. In the (almost)
horizontal part of the curve, around the voltage set point, the SVC control
is active. When the SVC has reached its maximum or minimum reactive
output, the voltage cannot be controlled, and the device will behave as a
pure reactor or pure capacitor; thus, in extreme voltage situations, the SVC
behaves as a reactor or capacitor bank. The control of the reactor current
is based on thyristors, which limits the bandwidth of the voltage control. If
a fast control is needed to compensate for ﬂicker and voltage dips, one has
to use technologies based on voltage source converters.
i
s
V
Figure 6.8. Static var compensator.
6.3. Primary Voltage Control 97
Figure 6.9. Static var compensator control.
STATCOM
The STATCOM is a device which is based on a voltage source converter.
The device is connected in shunt and consists of a capacitor charged with a
dc voltage, which provides the input voltage for a voltage source converter,
as illustrated in Figure 6.10. The converter feeds a reactive current into the
network, and by controlling this reactive current, voltage control is achieved.
Since a voltage source converter needs semiconductor elements with current
interrupting capabilities, thyristors cannot be used; instead, elements such
as GTOs or IGBTs have to be used.
In comparison with the SVC, the STATCOM oﬀers two advantages.
First, the STATCOM’s output reactive current is not limited at low or
highvoltage conditions; rather, the output current is only limited by the
converter ratings and is not dependent on the system voltage. This means
that the reactive support during extreme voltage situations is much bet
ter with respect to the SVC, as shown in Figure 6.11. Second, the control
response is much faster, since it is limited by the switching frequency of
the voltage source converter (usually around 1 kHz). The STATCOM can
hence be used to reduce ﬂicker and other fast voltage variations eﬀectively.
The dc capacitor of the voltage source converter constitutes an active power
storage; hence, an active current can also be injected into the network. The
STATCOM can thus also be used for active power control (example, to damp
power oscillations). However, the energy stored in the dc capacitor is fairly
small; therefore, to truly control the active power output, a battery must be
installed on the dc side.
98 6. Voltage Control in Power Systems
i
s
V
Voltage
source
converter
Figure 6.10. STATCOM.
Capacitive Inductive
i
s
V
Figure 6.11. Voltage control and reactive capability of a STATCOM.
6.4. Secondary Voltage Control 99
6.4 Secondary Voltage Control
As previously discussed, balancing the reactive power via AVRs and other
devices can be considered a local control action. Hence in case of a distur
bance, the devices electrically nearby will try to compensate the reactive
power needs. This may result in unacceptable voltage values and an uneven
reactive power distribution over the generators. This creates the need of a
coordinated adjustment of the setpoints of the reactive power suppliers.
Secondary Voltage Regulation (SVR) has been developed to address the
above mentioned situation. SVR is a hierarchical, centralized voltage control
scheme that supervises the generator AVRs and other reactive power sources
in a given network zone so as to enhance voltage stability of the grid. Figure
6.12 illustrates the basic setup. In each zone, a ’pilot’ node is selected and
controlled typically by the participating generators so as to maintain the
voltage of the pilot node at a speciﬁc value, by delivering the reactive power
proportionally to their own capabilities.
Powernetwork
AVR
QR
AVR
QR
AVR
QR
AVR
QR
AVR
QR
PI PI
TVR
SVR SVR
Figure 6.12. Structure of the voltage regulation.
Therefore the AVR is enhanced with two additional control levels. The
higher level is responsible for the voltage control of the ’pilot’ node. The
voltage error serves as input to a proportionalintegral controller (PI) and
the generated signal is then used by each generator at a lower control level.
The latter corresponds to a local reactive power regulator (QR) that adjusts
the input of the AVR with respect to the generator capability. A similar
hierarchy applies also in the time constants of each control level in order to
100 6. Voltage Control in Power Systems
avoid any interaction among them. The AVR time constant is in the order
of 0.5 sec, the QR time constant is in the order of 5 sec, whereas the time
constant of the PI controller is in the order of 50 sec.
The scheme can be further enhanced with the use of Tertiary Voltage
Regulation (TVR) which, based on an overall system economic optimization,
will determine the setpoints of the above mentioned PI controller in a time
scale of 15 minutes.
The described secondary closedloop control scheme is already imple
mented in some European power grids, particularly in France and in Italy
(since the early 1980’s). Other countries still follow an openloop strat
egy, in which the operator adjusts the setpoints manually with an expected
performance degradation.
7
Stability of Power Systems
In this chapter, factors inﬂuencing the stability of electric power systems
are discussed. First, a short introduction to damping in a power system is
given. Diﬀerent sources of positive and negative damping are discussed, and
methods to improve the damping are given. Furthermore, the eﬀect of the
loads on system stability is regarded in the second part of this chapter.
7.1 Damping in Power Systems
7.1.1 General
What damping in the context of electro–mechanical oscillations in a power
system means is quite self–evident. Normally, two diﬀerent kinds of electri
cal torques appear at a generator rotor that is oscillating: a synchronizing
torque ∆T
s
and a damping torque ∆T
d
. The synchronizing torque ∆T
s
is
in phase with the deviation in rotor angle ∆θ, whereas the damping torque
∆T
d
is in phase with the deviation in rotor speed ∆ω. The synchronizing
torque, also called the synchronizing power, strives, if it is positive, to bring
the rotor back to the stable equilibrium in which the mechanical power is
equal to the electrical power. When the generator has reached an operating
point where the synchronizing power no longer can return the system to
the stable equilibrium, the generator will fall out of phase. The variation
of the synchronizing torque with the rotor angle determines, together with
the machine’s moment of inertia, the frequency of rotor oscillations. The
partitioning into synchronizing and damping torque is shown in Figure 7.1.
Damping is neglected in the classical model. Therefore, the system will,
after a disturbance, either fall out of phase (instability) or oscillate with un
changed amplitude. This is not realistic, since real systems contain damping.
The damping torque depends on the time derivative of the rotor angle in
such a way that the oscillation is damped. Normally, the damping torque is
rather small and thus inﬂuences the oscillation frequency only marginally.
It mainly inﬂuences the amplitude. In a synchronous machine, the main
contributors to damping are the damper windings and the ﬁeld winding.
If the modes of oscillation in a system are determined by computing
the eigenvalues of the linearised system’s Jacobian matrix, changes in the
synchronizing and damping torques will become apparent as follows: An
increase of the synchronizing torque moves the eigenvalue parallel to the
101
102 7. Stability of Power Systems
∆ω
∆T
e
∆θ
∆T
s
∆T
d
Figure 7.1. Partitioning of electrical torque in synchronizing and
damping components.
imaginary axis towards larger values. This corresponds to an increase in the
spring constant in a mechanical analogy. If the damping torque is increased
instead, the eigenvalue will move parallel to the real axis to the left. In
the classical model, all eigenvalues will be situated on the imaginary axis.
Necessary for stability is that no eigenvalues are situated in the right half
plane. This corresponds to positive ∆T
s
as well as positive ∆T
d
in Figure 7.1.
7.1.2 Causes of Damping
As mentioned earlier, the internal damping of a generator comes from the
windings in the rotor circuit. That damping is determined by phases and
amplitudes of the oscillating torques caused by induced currents in exciter
winding and damping windings. Further, some loads contribute with posi
tive damping. These contributions originate from the frequency dependency
of the loads, but also their voltage dependency contributes.
Generally, the inner damping of the generators decreases with decreasing
frequency of the oscillations. The currents in the damping windings decay,
and hence, for very slow oscillations, their contribution is small.
Low or negative damping in a power system can lead to spontaneous ap
pearance of large power oscillations. This can, in the worst case, necessitate
tripping of lines and it must be avoided. That type of instability is called
small signal instability or instability caused by low damping. (Earlier, that
type of instability was called dynamic instability.)
A common reason for low damping is the use of voltage controllers with
high gain. That was experienced in generators feeding a strong net through
a line. Such a conﬁguration can also be analyzed comparably easily. It can
be shown that an eigenvalue with positive real part can occur when large
amounts of power are transmitted and voltage controllers with high gains are
used. Before the reason behind this phenomenon was known, the problem
was solved by operating the generator with manual voltage control, or by
7.1. Damping in Power Systems 103
K
STAB
sT
W
1 sT
W
+

1 sT
1
+
1 sT
2
+

∆ω
∆U
ref
Figure 7.2. Block Diagram of a simple PSS.
making the voltage controller slower or decreasing its gain.
To explain that mechanism in detail is beyond the scope of this com
pendium, but in summary, it can be said that the rotor angle inﬂuences
the generator voltage, which through the voltage controller inﬂuences the
transient emf, which inﬂuences the electrical torque. Now it turns out that,
when the load on the machine is high, the phase angle can be such that a
contribution with negative damping is obtained. If the ampliﬁcation in the
voltage controller is high, that negative contribution can be signiﬁcant.
7.1.3 Methods to Increase Damping
Several methods for increasing the damping in a power system are available.
The simplest and usually cheapest way is the installation of power system
stabilizers, PSS, in the generators. The operating principle for these is very
simple. To increase the damping in the system, a signal is added to the
reference voltage of the generator’s voltage controller. The phase of this
signal should of course be such that it results in a positive contribution to
the damping. Thus, the same physical mechanism in the system of generator
and voltage controller that above resulted in negative damping is used to
obtain positive damping.
Such a power system stabilizer usually utilizes the rotor deviation from
the synchronous frequency ∆ω as input signal. Sometimes, other signals
that contain the same information can be used, like P
e
or T
e
. A diagram
illustrating the principle mode of operation of a PSS is given in Figure 7.2.
The input signal, in this case ∆ω, ﬁrst passes a high–pass ﬁlter to ensure that
permanent frequency deviations do not contribute. The next ﬁlter shifts the
phase appropriately for the critical oscillation frequency so that a positive
contribution to damping is obtained. The constant K
stab
determines the
size of that contribution. That constant should of course not be chosen
larger than necessary to obtain the needed damping, since this could lead
to undesired side eﬀects.
Other possibilities for increasing the damping in a system are diﬀerent
types of controllable equipment that may be installed in the system, such as
HVDC (High Voltage Direct Current) or SVC (Static Voltage Condensers).
These components can often give large contributions to damping, but they
are usually too expensive to install them only to increase the damping, and
the existing equipment is not always located optimally for damping purposes.
104 7. Stability of Power Systems
7.2 Load Modelling
Since, neglecting losses, an equal amount of power is consumed in the loads
in the system as is generated in the generators, the load characteristics are on
principle just as important for the system properties as the generators. That
is, however, not reﬂected in the level of detail and the accuracy usually used
in load models for analyzing system stability. This chapter discusses brieﬂy
how load characteristics inﬂuence the system stability and which problems
arise in the derivation of appropriate load models. The most common load
models are presented.
7.2.1 The Importance of the Loads for System Stability
The characteristics of the loads inﬂuence the system stability and dynamics
in many diﬀerent ways. The voltage characteristics of the loads have a
direct inﬂuence on the accelerating power for generators nearby and are
thus very important for the behaviour during the ﬁrst oscillation after a
fault. It has been shown in section 2.1.2 that the frequency dependency of
the loads inﬂuences directly how large the frequency deviation after diﬀerent
system disturbances will become. The frequency dependency of the loads
also inﬂuences the system damping. The same is true for their voltage
dependency since it inﬂuences the voltage control.
This compendium concentrates on what is usually called angular stabil
ity, or synchronous stability, that is, the ability of the generators to stay
synchronized after disturbances. Another important property of a power
system is the ability to keep the voltages in the system within acceptable
limits during disturbances. This is a measure for the voltage stability of
the system. Voltage stability is highly dependent on the balance of reactive
power in the system, but also the active power has some inﬂuence here. It
is obvious that the voltage dependency of the loads is of high importance
for the system’s voltage stability.
It is for several reasons diﬃcult to derive good load models. (Of course,
deriving models for single load objects is formally not very diﬃcult. Loads
here are, however, lumped loads as they are perceived from a bus in the
high voltage grid.) First, it is diﬃcult to estimate the composition of the
loads, since it varies during the day as well as during the year. Further,
this composition varies from bus to bus. Thus, sometimes diﬀerent load
models have to be used at diﬀerent buses, depending on the composition of
the loads, for example industrial loads, domestic loads, and rural loads.
7.2.2 Load Models
For studies of angular stability, loads are usually modelled with static mod
els. Sometimes, large induction motors have to be represented individually
7.2. Load Modelling 105
by special models to obtain the correct dynamic behaviour. Dynamic load
models for lumped loads have begun to be used during the last few years,
especially for studying voltage stability, but those are expected to be used
in the future more widely and even for other types of studies.
Static Load Models
For traditional stability studies, where the investigated time frame is at
most around 10 s after the disturbance, the most commonly used model
types are static models. They are called static since they describe the load
using only algebraic equations. The modelled load dynamics are in these
cases so fast that they can be considered instantaneous compared with other
phenomena, like rotor oscillations, that are modelled. The most common
model for voltage dependency is
P = P
0
_
U
U
0
_
α
,
Q = Q
0
_
U
U
0
_
β
.
(7.1)
However, the load can also be modelled as an arbitrary polynomial in
(U/U
0
),
P = P
0
i
k
i
_
U
U
0
_
α
i
. (7.2)
In Equations (7.1) and (7.2), U
0
is the nominal voltage at nominal load,
P
0
and Q
0
. In Equations (7.1), the voltage exponents α and β are often
diﬀerent.
If α = 0, the load is called constant power load; if α = 1, it is called
constant current load; and if α = 2, it is a constant impedance load. It is
very common to model the (active) load for stability studies as consisting of
these three parts. Usually, a somewhat larger voltage exponent is used for
the reactive load.
Some examples for the voltage exponents of diﬀerent loads are:
Electric heating: α = 2, Q = 0.
Light bulbs: α ≈ 1, 6, Q = 0.
Fluorescent tubes: α ≈ 0, 9, β ≈ 2.
To be able to include active and reactive losses in the underlying distribution
grid, the above models have to be modiﬁed.
A load model often used is the so called ZIP–Model. The ZIPModel is
a special case of the model in Equation (7.2), which contains three terms
with:
• α
1
= 2, i.e. constant impedance (Z) load
106 7. Stability of Power Systems
r
s
i
e′
+
e
+
X′ ωL′ =
Figure 7.3. Representation of induction motor.
r
s
L
s
L
r
L
m
r
r
s

Figure 7.4. Equivalent circuit for induction motor.
• α
2
= 1, i.e. constant current (I) load
• α
3
= 0, constant power (P) load
Motor Loads
Around half of all electric power used by the industry is used for operation
of motors. Sometimes, the load in certain nodes is dominated by electric
motors. It can then be justiﬁed to model those explicitly.
For small changes in voltage, a motor load behaves approximately like a
constant power load. For larger voltage changes, it can be necessary to use
a more accurate representation. Synchronous machines are then modelled
according to the models derived in Chapter 5, with the mechanical part P
m
depending on the characteristic of the mechanical load. A large part of the
motor load consists of induction motors that can be modelled as follows:
An induction motor is basically a synchronous machine with short–circuited
exciter coil. If the exciter coil rotates with an angular speed diﬀerent from
the rotating ﬂuxes generated by the three phase coils, a current that gener
ates a ﬂux is induced in the exciter coil. Between the rotating synchronous
ﬂux generated in the phase windings and the ﬂux from the exciter wind
ing, energy is exchanged. This is the basis for the function of the induction
motor.
The induction motor can, according to Figure 7.3, be described by a
voltage source behind an impedance. The value of L
′
can be obtained from
the equivalent circuit of the induction motor shown in Figure 7.4.
7.2. Load Modelling 107
In Figure 7.4
r
s
and L
s
are the stator resistance and inductance,
L
m
is the magnetizing inductance,
r
r
and L
r
are rotor resistance and inductance.
The slip s is deﬁned by
s =
ω
0
−ω
ω
0
(7.3)
and thus L
′
is given by
L
′
= L
s
+
L
m
L
r
L
m
+L
r
. (7.4)
The dynamics are described by
de
′
dt
+
1
τ
′
0
(e
′
+jω
0
(L
s
+L
m
−L
′
)i) +jω
0
se
′
= 0 , (7.5)
i =
e −e
′
r
s
+jω
0
L
′
, (7.6)
dω
dt
=
1
2H
m
(T
e
−T
l
) , (7.7)
T
e
= ℜ(e
′
· i
∗
) . (7.8)
Here,
ω is the machine’s angular speed,
ω
0
is the system’s angular speed,
T
l
is the load torque,
τ
′
0
= (L
r
+L
m
)/r
r
are no load operation constants.
Equivalent Dynamic Loads
The load models presented above are, as mentioned, valid for studying phe
nomena that do not last longer than about ten seconds after a disturbance.
If phenomena taking place in a longer time frame should be studied, slow
dynamics in the system have to be accounted for. These dynamics origi
nate mainly from two diﬀerent sources: The tap changers installed at lower
voltage levels that try to restore the voltage to the desired value and the
controllers installed at the loads.
The control of tap changers can be done in several diﬀerent ways, but
common to most systems are that tap changers are stepped, typically in
intervals of some tens of seconds, until the voltage is restored . Since this
control exists at diﬀerent voltage levels (cascade coupled controllers), un
desirable overshoots in the control can occur if the control loops are not
coordinated. Generally, the control has to be slower the lower the voltage
level is.
108 7. Stability of Power Systems
In Sweden, a large part of the load, at least in winter, consists in many
areas of heating loads. The changes in this type of load are determined by
thermostats. Hence, it takes some time until, for example, a voltage drop
becomes apparent. That time is determined by the thermal time constants
for what is heated, such as houses, and by the design of the thermostats.
Summarizing, it can be said that the dynamics determined by tap–chan
ger control and load dynamics are highly complicated. Measurements are
needed to get reliable results. Measurements of load characteristics have
during the last few years become very important, and much work is be
ing done in many utilities to investigate load characteristics under diﬀerent
loading conditions.
A typical example of a load behaviour after a voltage drop is shown in
Figure 7.5. It is clearly visible how the load drops momentarily, as described
by the load models from Section 7.2.2, to recover later to a considerably
higher level. A rather general description is given by
T
p
dP
r
dt
+P
r
= P
s
(U) −P
t
(U) , (7.9)
P
l
(t) = P
r
+P
t
(U) , (7.10)
where
P
r
(t) is a state variable,
P
s
(U) is a static model for the long term load behaviour,
P
t
(U) is a static model for the transient load behaviour,
P
l
(t) is the value of the active load at the time t.
Of course, U = U(t) in the equations above.
For the reactive load, similar behaviour and equations are valid.
7.2. Load Modelling 109
t
P ∆
P
s
∆
t
P ∆
P
s
∆
t
P ∆

PRF
=
20 15 10 5 0 5
0.75
0.8
0.85
0.9
0.95
1
1.05
Time [minutes]
A
c
t
i
v
e
P
o
w
e
r
[
p
u
]
Load Recovery Factor Definition
Figure 7.5. The Transient Behaviour of the Load in Equations (7.9)
and (7.10) after a Step in the Voltage.
110 7. Stability of Power Systems
References
[1] P. Kundur: Power System Stability and Control, McGrawHill Inc.,
New York, 1994 (ISBN 007035958X).
[2] Dynamic Models for Steam and Hydro Turbines in Power System Stud
ies, IEEE Trans. Power Appar. Syst. 19041915, Nov./Dec. 1973.
[3] G. Andersson: Modelling and Analysis of Electric Power Systems, ETH
Zurich, 2009.
[4] W.G. Heﬀron and P.A. Phillips: Eﬀect of modern aplidyne voltage
regulator on underexcited operation of large turbine generators, Trans.
Am. Inst. Electr. Eng., Part 3, 71, 692 – 697, 1952. As cited in: YaoNan
Yu, Electric Power System Dynamics, Academic Press, 1983.
[5] A.R. Bergen and V. Vittal: Power System Analysis, Second Edition,
Prentice Hall, 2000.
[6] A. GomezExposito, A. J. Conejo and C. Canizares: Electric Energy
Systems: Analysis and Operation, CRC Press, 2009.
111
112 REFERENCES
Appendix A
Connection between per unit and SI
Units for the Swing Equation
If nothing else is given after a quantity, that quantity is in SI–units. If a
quantity is expressed in per unit, p.u. is given in brackets after the quan
tity (p.u.). For simplicity, it is assumed that the nominal electrical and
mechanical frequencies are equal, (rad/s).
In SI–units,
J
d
2
θ
dt
2
= ∆T , (A.1)
with
J = moment of inertia for rotor turbine (kgm
2
),
θ = angle (rad),
ω = angular velocity (rad/s),
∆T = eﬀective torque on the rotor turbine (Nm).
When using electrical degrees, Equation (A.1) is usually written as
M
ω
0
·
d
2
θ
dt
2
= ∆T , (A.2)
with
M = moment of inertia = Jω
0
180
π
(Js/el
◦
).
The Hfactor, or constant of inertia, for synchronous machine i is deﬁned
by
H
i
=
1
2
Jω
2
0
S
i
, (A.3)
with
H
i
= constant of inertia for synchronous machine i (s),
S
i
= rated power of synchronous machine.
The per unit base for torques at synchronous machine i, T
bas,i
, is given by
T
bas,i
=
S
i
ω
0
, (A.4)
leading to
∆T = ∆T(p.u.)
S
i
ω
0
. (A.5)
Using (A.3) and (A.5), (A.1) can be written as
2H
i
S
i
ω
2
0
·
d
2
θ
dt
2
=
S
i
ω
0
∆T(p.u) , (A.6)
113
114 A. Connection between per unit and SI Units for the Swing Equation
or
d
2
θ
dt
2
=
ω
0
2H
i
∆T(p.u) . (A.7)
Equation (A.7) can also be written as
d
dt
(
˙
θ) =
ω
0
2H
i
∆T(p.u) , (A.8)
which is the same as
d
dt
(ω) =
ω
0
2H
i
∆T(p.u) , (A.9)
or
d
dt
(ω(p.u.)) =
∆T(p.u)
2H
i
. (A.10)
Generally,
P = Tω
m
, (A.11)
with the actual mechanical angular speed of the rotor ω
m
that, according
to the assumptions, is equal to the electrical angular speed ω. With the
equations above, this gives
P(p.u) = T(p.u) ω(p.u) . (A.12)
Equation (A.10) now becomes
˙ ω(p.u.) =
∆P(p.u.)
2H
i
·
1
ω(p.u.)
. (A.13)
For rotor oscillations, ω(p.u.) ≈ 1 and (A.13) can be approximated by
˙ ω(p.u.) =
∆P(p.u.)
2H
i
, (A.14)
or
˙ ω =
ω
0
2H
i
∆P(p.u.) . (A.15)
The most common equations in literature are (A.9), (A.10), (A.14), and
(A.15). Of these, (A.9) and (A.10) are exact if ω is the actual angular
frequency. Equations (A.14) and (A.15) are good approximations as long as
ω ≈ ω
0
. That is valid for “normal” oscillations in power systems.
Appendix B
Inﬂuence of Rotor Oscillations on the
Curve Shape
If the relative movement between the ﬁeld winding of a synchronous machine
and its phase windings is a purely rotating motion with constant angular
speed, the resulting induced voltages in the phase windings will be shaped
ideally like a sinusoid. From now on, it is assumed that the ﬁeld winding is in
the rotor, while the phase windings are on the stator, but since the relative
motion determines the voltage in the phase windings, it is even possible
to think of stationary ﬁeld windings and rotating phase windings. In all
modern larger synchronous machines, the ﬁeld winding is on the rotor, so
the assumption above does have a practical background. However, almost
all relationships and conclusions are independent of this assumption.
For simplicity, consider a single phase synchronous machine according
to Figure B.1. A three phase machine has two more phase windings shifted
±120
◦
relative to the phase winding in the ﬁgure. The phase winding and
the exciter winding are arranged so that the ﬂux linkage through the phase
winding is sinusoidally shaped as a function of the angle θ
m
in Figure B.1:
Φ(t) = Φ
0
cos θ
m
= Φ
0
cos ω
0
t , (B.1)
with the angular speed of the rotor ω
0
according to the system’s electrical
frequency. That ﬂux induces a voltage in the phase winding that is given by
U(t) =
dΦ
dt
= −Φ
0
ω
0
sinω
0
t = −
ˆ
U sin ω
0
t . (B.2)
Now, we shall study how the ﬂux linkage through the phase windings will
be inﬂuenced when rotor oscillations appear in the system. If the balance
between power into the generator and power from the generator, i.e. between
mechanical torque and electrical power, is disturbed, the rotor will start to
oscillate relative to an undisturbed reference rotor that continues to rotate
with the angular speed ω
0
. The rotor position can generally be described
by
θ
m
(t) = ω
0
t +θ(t) , (B.3)
where θ(t) is a solution of the swing equation. It has earlier been mentioned
that stable solutions of the swing equation for a synchronous machine con
nected to a strong grid consist of oscillations that are nearly sinusoidal with
frequencies on the order of magnitude of some tenths of a Hertz to some
115
116 B. Inﬂuence of Rotor Oscillations on the Curve Shape
θ
m
S
N
Figure B.1. Schematic picture of single phase synchronous machine.
Hertz. To investigate how the linked ﬂux, and thus the voltage and the
current, look during an oscillatory movement, a rotor motion according to
θ
m
(t) = ω
0
t +µsin(ω
r
t +θ
r
) , (B.4)
with the angular speed ω
r
corresponding to the oscillation frequency and
the amplitude of the oscillatory movement µ, is assumed. The ﬂux linkage
can now be written as
Φ(t) = Φ
0
cos θ
m
(t) = Φ
0
cos(ω
0
t +µsin(ω
r
t +θ
r
)) , (B.5)
which implies that the oscillatory movement contains a phase–angle modu
lation of the ﬂux linkage. The momentary angular frequency, Ω(t), is deﬁned
for Ψ(t) as
Ω(t) =
d
dt
(ω
0
t +µsin(ω
r
t +θ
r
)) (B.6)
and varies between ω
0
+µω
r
and ω
0
−µω
r
. It can be shown (cf. text books
on modulation theory) that Equation (B.5) can be written as
Φ(t) = Φ
0
n=∞
n=−∞
J
n
(µ) cos((ω
0
+nω
r
)t +nθ
r
) . (B.7)
J
n
(µ) is a Bessel function of the ﬁrst kind with the argument µ and degree
n, as given by
J
n
(µ) =
1
π
_
π
−π
cos(µsin x −nx)dx . (B.8)
An important property of J
n
(µ) that will be used later is
J
−n
(µ) = (−1)
n
J
n
(µ) . (B.9)
117
J ( ) µ
3
J ( ) µ
2
J ( ) µ
1
J
0
( ) µ
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0
µ
Figure B.2. Bessel functions of the ﬁrst kind of order 0 to 3.
From Equation (B.7) follows that, in spite of that the momentary angular
frequency for Φ(t) is between ω
0
+µω
r
and ω
0
−µω
r
, Φ(t) will have inﬁnitely
many side bands with the frequencies ω
0
± nω
r
beside the fundamental
frequency ω
0
. A relevant question is how large the amplitudes of these
side bands are. Generally, the coeﬃcients J
n
(µ) decay rapidly when the
order n becomes larger than the argument µ. In Figure B.2, values for
the ﬁrst four Bessel functions are shown for the argument µ between 0
and 1. It should be observed that µ is measured in radians, so that 1
corresponds to approximately 57
◦
, which in this context is quite a large
amplitude. Figure B.2 shows that side bands with n = 3 and larger can
be neglected even for amplitudes as large as µ = 1. Φ(t) can thus be
approximated quite accurately by
Φ(t) ≈ Φ
0
n=2
n=−2
J
n
(µ) cos((ω
0
+nω
r
)t +nθ
r
) . (B.10)
Typical oscillation frequencies are, in most cases considerably, lower than
3 Hz, so that practically the whole energy spectrum for Φ(t) lies in the
frequency area f
0
± 6 Hz, i.e. 50(60) ± 6 Hz. This justiﬁes the usual
representation of the grid with the traditional phasor model with a constant
frequency corresponding to f
0
.
If the amplitude is very small, say less than µ ≈ 0.2, corresponding to
an amplitude of approximately 10
◦
, side bands with n = 2 and higher can
118 B. Inﬂuence of Rotor Oscillations on the Curve Shape
ω
r
ω
r
ω
r
µ
2

µ
2

1
Figure B.3. Vector with superimposed oscillation, according to the text.
be neglected, see Figure B.2. Further, it can be shown that, if µ is small,
J
0
(µ) ≈ 1 −
µ
2
4
,
J
1
(µ) ≈
µ
2
,
(B.11)
are valid, and Φ(t) can be written as
Φ(t) ≈ Φ
0
((1−
µ
2
4
) cos ω
0
t+
µ
2
cos((ω
0
+ω
r
)t+θ
r
)−
µ
2
cos((ω
0
−ω
r
)t−θ
r
)) .
(B.12)
Equation (B.9) has here been used. That approximation is used even in a
context of power system analysis other than rotor oscillations, namely when
studying so–called subsynchronous oscillations, SSO. If there is a resonance
phenomenon at a subsynchronous frequency it is called subsynchronous res
onance, SSR. The most common cause of SSO are torsional oscillations on
the axis connecting the turbine(s) and the generator rotor. The frequencies
of the natural oscillations on that axis are typically 5 Hz and higher. (For
an axis with n distinct “masses”, including generator rotor and, maybe,
exciter, there are n − 1 diﬀerent eigenfrequencies.) For the side band in
Equation (B.12), i.e. for ω
0
± ω
r
, the frequency can deviate signiﬁcantly
from the nominal frequency, so that it is normally not possible to look only at
the component with nominal frequency. Since the electrical damping for the
lower side band, the subsynchronous frequency, can be negative, due to, for
example, series compensation, the partitioning according to Equation (B.12)
has to be kept. The damping in the upper side band, the supersynchronous
frequency, is almost always positive.
To increase the understanding for the partitioning in Equation (B.12), a
more intuitive derivation than the stringent mathematical one using Bessel
119
functions can be given. Consider a vector with amplitude 1 that performs
small oscillations with an angular frequency ω
r
and amplitude µ. This can
be illustrated geometrically according to Figure B.3. The vector with ﬁlled
arrowhead oscillates symmetrically around the horizontal axis with the fre
quency ω
r
and the amplitude µ. That vector can now be partitioned into
the three vectors with unﬁlled arrowheads. One vector does not move and
lies along the horizontal axis. Two vectors with amplitude µ/2 rotate with
the angular frequency ±ω
r
according to Figure B.3. It is easily observed
that the sum of the vectors with hollow arrowhead is at all times equal to
the vector with ﬁlled arrowhead. Since the vectors rotate with the angular
frequency ω
0
with respect to a stationary system, Equation (B.12) is ob
tained directly from the projection of the vectors on to the horizontal axis,
with the modiﬁcation that the factor for the fundamental frequency is one.
ii
Contents
Preface 1 Introduction 1.1 Control Theory Basics  A Review 1.1.1 Simple Control Loop . . . . 1.1.2 State Space Formulation . . 1.2 Control of Electric Power Systems 1.2.1 General considerations . . . v 1 2 2 5 5 5 9 9 9 14 17 18 18 21 21 23 24 27 27 29 31 33 36 36 45 46 47
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
2 Frequency Dynamics in Electric Power Systems 2.1 Dynamic Model of the System Frequency . . . . . 2.1.1 Dynamics of the Generators . . . . . . . . . 2.1.2 Frequency Dependency of the Loads . . . . 2.2 Dynamic Response of Uncontrolled Power System . 2.3 The Importance of a Constant System Frequency . 2.4 Control Structures for Frequency Control . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
3 Primary Frequency Control 3.1 Implementation of Primary Control in the Power Plant . . . . 3.2 Static Characteristics of Primary Control . . . . . . . . . . . 3.2.1 Role of speed droop depending on type of power system 3.3 Dynamic Characteristics of Primary Control . . . . . . . . . . 3.3.1 Dynamic Model of a OneArea System . . . . . . . . . 3.3.2 Dynamic Response of the OneArea System . . . . . . 3.3.3 Extension to a TwoArea System . . . . . . . . . . . . 3.3.4 Dynamic Response of the TwoArea System . . . . . . 3.4 Turbine Modelling and Control . . . . . . . . . . . . . . . . . 3.4.1 Turbine Models . . . . . . . . . . . . . . . . . . . . . . 3.4.2 Steam Turbine Control Valves . . . . . . . . . . . . . 3.4.3 Hydro Turbine Governors . . . . . . . . . . . . . . . . 3.5 Dynamic Responses including Turbine Dynamics . . . . . . . iii
. . . . . . . A Connection between per unit and SI Units for the Swing Equation 113 B Inﬂuence of Rotor Oscillations on the Curve Shape 115 . . . . . . .1. . . . . . . . . . . . . . . . . . . . . . . . .3. . . 5. . . . . . . . .2.2 The Inductance Matrices of the Synchronous Machine . . . . . . . . . .4 Secondary Voltage Control . . . . . . . . .1 Static Characteristics of AGC . . . . . . 6. . . . 7. . . . .1 Synchronous Machine Excitation System and AVR 6. . . . . . . . . . . .2.3 Primary Voltage Control . . .6. . . . . . . . . . . . . . . . . .1 General . .3 Twoarea system. . . . . . .5 Time constants .1 Onearea system . . . . . . . .4 Synchronous. . . .1 Derivation of the fourthorder model . 4. . . . . . . . . . . .1 Park’s Transformation . . . . . . . . . . . . . . . . . . . 6. . . 5 Synchronous Machine Model 5. . . . .3 Transformer Tap Changer Control . .2. . . . 4. . . . . . . . . . 6. . .1 The Importance of the Loads for System Stability 7.2 The HeﬀronPhillips formulation for stability studies 6 Voltage Control in Power Systems 6. . . . . . . . 6. . . . . . . . .2 Voltage Control Mechanisms . . . . . . . . including saturations – disturbance response . . . . . . .2 Load Models . . . . . . .3 Methods to Increase Damping . . . . . . . . . . . . . . . . . . . .1. . . . . . . . . . . . 6. . .2 Reactive Shunt Devices .6. 7. .4 Twoarea system – equal sizes. . . . . . . . . . . . 5. 5. . . . . . . . . . . . . . . . . . . . . . . . . 5. .1 Damping in Power Systems . . . . . . . . . . . . . . . . . . . . . . 7. .2.6 Simpliﬁed Models of the Synchronous Machine . . . . 5. . Transient. . . . . . . . . . and Subtransient Inductances . 6. . . . . . . . . . . . . . . . . . 5. . . . . . . . .2 Twoarea system – unequal sizes – disturbance response 4. . .iv Contents 51 51 54 54 56 58 60 61 61 65 67 70 74 76 76 79 85 85 87 88 88 93 94 95 99 101 101 101 102 103 104 104 104 110 4 Load Frequency Control 4.4 FACTS Controllers . . . . .1. . . . . . . 7. . . .2 Load Modelling . References . . . 4.2. . . . . . . . . . .1 Relation between voltage and reactive power .2. . unequal sizes – normal control operation . . . . . 5. .3 Voltage Equations for the Synchronous Machine . . . 7. . . . . . . . . . .2 Dynamic Characteristics of AGC . . . . . 7 Stability of Power Systems 7. . . . .3. . .3. . . . . . . . . 4. . . .2 Causes of Damping . . .3. .
is also included. ETH Z¨rich) given at u ETH Z¨rich in the Master Programme of Electrical Engineering and Inforu mation Technology. Z¨rich. The main topic covered is frequency control in power systems.Preface These lectures notes are intended to be used in the lecture Dynamics and Control of Power Systems (Systemdynamik und Leittechnik der elektrischen Energieversorgung) (Lecture 227052800. DITET. The needed models are derived and the primary and secondary frequency control are studied. An overview of load models is also given. February 2012 u G¨ran Andersson o v . The excitation and voltage control of synchronous machines are brieﬂy described. based on Park’s transformation. A detailed model of the synchronous machine.
vi Preface .
This means that the control system consists of a number of nested control loops that control or regulate diﬀerent quantities in the system. operates in a time scale of tens of seconds or minutes. which determines the reference values of the voltage controlling devices. In general the control loops on lower system levels.1. 1 . As an example. while the Secondary Voltage Control. and Consumption (Loads). which regulates the voltage of the generator terminals to the reference (set) value. responds typically in a time scale of a second or less. Some basic results from control theory are reviewed.g. and an overview of the use of diﬀerent kinds of power plants in a system is given. The studied system comprises the subsystems Electricity Generation. The main topics of these lectures will be • Power system dynamics • Power system control • Security and operational eﬃciency. That means that these two control loops are virtually decoupled. are characterized by smaller time constants than the control loops active on a higher system level.1 Introduction In this chapter a general introduction to power systems control is given. A schematic diagram showing the diﬀerent time scales is shown in Figure 1. e. Distribution. resulting in a number of decoupled control loops operating in diﬀerent time scales. locally in a generator. Transmission. among which the generators. the Automatic Voltage Regulator (AVR). This is also generally true for other controls in the systems. In order to study and discuss these issues the following tools are needed • Control theory (particularly for linear systems) • Modelling • Simulation • Communication technology. and the associated control system has a hierarchic structure.
This facilitates the task. e. i.e.1 Simple Control Loop The control system in Figure 1.3 is considered.1. is within speciﬁed limits. This consumption is normally uncontrolled. voltage magnitudes and frequency. the system should be operated in an economically eﬃcient way. could be disconnected due to faults must be taken into account. one usually has to resort to computer simulations. but due to the decoupling it is in most cases possible to study the diﬀerent control loops individually.A Review The decoupled control loops described above can be analyzed by standard methods from the control theory.1. Just to refresh some of these concepts. Furthermore. since substantial parts of the system are exposed to external disturbances. Schematic diagram of diﬀerent time scales of power system controls.2 Tie Line Power and Frequency Control Turbine Control Voltage Control Protection 1. In this ﬁgure the block G(s) represents the controlled plant and also possible controllers. 1. and to explain the notation to be used. 1. the electric power consumption. For a more detailed analysis. a very short review is given here.g. the possibility that lines etc.1 Control Theory Basics . The overall control system is very complex. This has resulted in a hierarchical control system structure as shown in Figure 1. In addition. The task of the diﬀerent control systems of the power system is to keep the power system within acceptable operating limits such that security is maintained and that the quality of supply. and with appropriate simpliﬁcations one can quite often use classical standard control theory methods to analyse these controllers. Introduction 1/10 1 10 100 Time (s) Figure 1.2. A characteristic of a power system is that the load. varies signiﬁcantly over the day and over the year. From this ﬁgure .
In the following this convention is not always adhered to. In principle two diﬀerent problems are solved in control theory: 1. Tracking problem Here the quantities in the time domain are denoted by small letters. Regulating problem 2. while the Laplace transformed corresponding quantities are denoted by capital letters.1. The structure of the hierarchical control systems of a power system. Control Theory Basics . 1 .1.A Review 3 System Control Center State Estimation Power Flow Control Economic Dispatch Security Assessment NETWORK Generation (In Power Stations) Turbine Control Voltage Control Reactive Power Compensation HVDC. the following quantities are deﬁned1 : • r(t) = Reference (set) value (input) • e(t) = Control error • y(t) = Controlled quantity (output) • v(t) = Disturbance Normally the controller is designed assuming that the disturbance is equal to zero.2. but it should be clear from the context if the quantity is expressed in the time or the s domain. FACTS Power Transmission And Distribution Tap Changer Control (Direct and Quadature) Loads Normally not controlled Figure 1. but to verify the robustness of the controller realistic values of v must be considered.
Simple control system with control signals. to the output. etc.4 1.2) (1. the initial and stationary response of the output would be y(t → 0+) = lim F (s) s→∞ (1. This is sometimes also called the servo problem.4) (1.5) and y(t → ∞) = lim F (s) s→0 . Introduction v r + e y G(s) _ Σ H(s) Figure 1. R.1) In many applications one is not primarily interested in the detailed time response of a quantity after a disturbance.3. Then the two following properties of the Laplace transform are important: g(t → 0+) = lim sG(s) s→∞ (1. The transfer function from the input. This is the most common problem in power systems. and F (s) is the transfer function. If the input is a step function. and other quantities should be kept at the desired values irrespective of load variations. but rather the value directly after the disturbance or the stationary value when all transients have decayed. the reference value r is normally kept constant P and the task is to keep the output close to the reference value even if disturbances occur in the system. frequency. Y . is given by (in Laplace transformed quantities) F (s) = Y (s) C(s) G(s) = = R(s) R(s) 1 + G(s)H(s) (1. where the voltage.3) and g(t → ∞) = lim sG(s) s→0 where G is the Laplace transform of g. In the tracking problem the task is to control the system so that the output y follows the time variation of the input r as good as possible. Laplace transform = 1/s. line switchings. In the regulating problem.
In most realistic cases D = 0. and of the latter Western and Southern USA. over the year. of dimension n × n. The matrix A. The matrices A and B deﬁne which states are controllable. where the matrix K deﬁnes the feedback control. The matrices B. so called winter peak. and D depend on the design of the controller and the available outputs. The consumption varies also over the day as shown in Figure 1. Control of Electric Power Systems 5 1. This means that the voltage magnitude.2. If a change in the load occurs. rotor and turbines. this will lead to frequency deviations that if too large will have serious impacts on the system operation.4. .2 State Space Formulation The vector x = (x1 x2 . which uniquely describe the system. The vector u has the inputs as components. In the long time scale.2 1. the controlled system becomes x = (A − BKC)x ˙ (1. and the vector y contains the outputs as components.7) A linear and timeinvariant controlled system is deﬁned by the equations ˙ x = Ax + Bu (1.4.6) y = Cx + Du 1. is the system matrix of the uncontrolled system. A complication is that the electric power consumption varies both in the short and in the long time scales.1 Control of Electric Power Systems General considerations The overall control task in an electric power system is to maintain the balance between the electric power produced by the generators and the power consumed by the loads. xn )T contains the states of the system. so called spontaneous load variations. the delivered electricity must conform to certain quality criteria.1. Examples of the former are most European countries. summer peak. and the matrices A and C deﬁne which states are observable. Also in the short run the load ﬂuctuates around the slower variations shown in Figure 1. which means that there is zero feedthrough.2. this is in the ﬁrst step compensated by the kinetic energy stored in the rotating parts. of the . at all time instants. A controller using the outputs as feedback signals can be written as u = −Ky = −KCx. . and wave shape must be controlled within speciﬁed limits. C. including the network losses. frequency.1. If this balance is not kept. and the system is said to be strictly proper. the peak loads of a day are in countries with cold and dark winters higher in the winter. assuming D = 0. while countries with very hot summers usually have their peak loads in summer time. In addition to keeping the above mentioned balance.
If several power producers are competing on the market. it usually takes several hours to get the unit operational. 2 . which is done through the frequency control of the generators in operation. Typical load variations over a day. An unbalance in the generated and consumed power could also occur as a consequence of that a generating unit is tripped due to a fault. Left: Commercial load. the situation is more complex. while for thermal power plants. Right: Residential load. conventional or nuclear.. liberalized electricity market. the power supplied from the generators must be changed. the unit commitment was made in such a way that the generating costs were minimized. The task of the frequency control is to keep the frequency deviations within acceptable limits during these events. If this frequency change is too large. Introduction 100 80 120 100 weekdays holidays holidays weekdays holidays Load (%) 60 40 20 0 1 4 7 10 Load (%) 80 60 40 20 0 13 16 19 22 1 4 7 10 13 16 19 22 Time Time Figure 1. This has an impact on the unit commitment and on the planning of reserves in the system3 . For hydro units and gas turbines this time is typically of the order of some minutes. To cope with the larger variations over the day and over the year generating units must be switched in and oﬀ according to needs. but also the time it takes to bring a generator online from a state of standstill. But also in these cases a unit commitment must be made. intermediate load.6 1. Such a plan is called unit commitment. This classiﬁcation is based on the time it takes to activate the plants and on the With the methods available today one can make a load forecast a day ahead which normally has an error that is less than a few percent. generators resulting in a frequency change. and a simple cost minimizing strategy could not be applied. etc. pool. Plans regarding which units should be on line during a day are done beforehand based on load forecasts 2 .4. 3 In a system where only one company is responsible for the power generation. but according to other principles. When making such a plan. economic factors are essential. The competing companies are then bidding into diﬀerent markets. Depending on how fast power plants can be dispatched. or base load power plants. they are classiﬁed as peak load. bilateral.
it is obvious that the power plant should be used during high load conditions when generating capacity is scarce. 4 . • Peak load units. operational time 3000–4000 h/a – Fossil fuel thermal power plants – Bio mass thermal power plants • Base load units. The overall goal of the unit commitment and the economic dispatch is the • Minimization of costs over the year • Minimization of fuel costs and start/stop costs The fuel costs should here be interpreted more as the “value” of the fuel. Control of Electric Power Systems 7 fuel costs and is usually done as below4 .5 the use of diﬀerent power plants is shown in a load duration curve representing one year’s operation. The classiﬁcation is not unique and might vary slightly from system to system. which can be interpreted as a high fuel cost. operational time 5000–6000 h/a – Run of river hydro power plants – Nuclear power plants In Figure 1. operational time 1000–2000 h/a – Hydro power plants with storage – Pumped storage hydro power plants – Gas turbine power plants • Intermediate load units. For a hydro power plant the “fuel” has of course no cost per se.2. But if the hydro plant has a storage with limited capacity. This means that the “water value” is high.1.
Duration curve showing the use of diﬀerent kinds of power plants. .8 1. Introduction System Load Gas turbines Hydro power reserves Thermal power (fossil fuel) Controllable hydro power Nuclear power Run of river hydro power Time 1 year Figure 1.5.
1 Dynamic Model of the System Frequency In order to design a frequency control methodology for power systems.1. like a loss of generation. a simpliﬁed model of a power system with several generators (synchronous machines) will be derived in the sequel. The frequencies of the diﬀerent machines can be regarded as comparatively small variations over an average frequency in the system. Certain simpliﬁcations lead to a description of the dominant frequency dynamics by only one diﬀerential equation which can be used for the design of controllers. the frequency in diﬀerent parts of a large power system will vary similar to the exemplary illustration shown in Figure 2. It is based on the swing equation for the set of synchronous machines in the system. the elementary dynamic characteristics of the system frequency have to be understood.2 Frequency Dynamics in Electric Power Systems In this chapter. This average frequency.1. For this purpose. Note that no control equipment is present yet in the models presented in this chapter: the system is shown in ”open loop” in order to understand the principal dynamic behaviour. called the system frequency. 2. For this purpose. Control methods are presented in the subsequent chapters. the basic dynamic frequency model for a large power system is introduced. which has an accelerating or decelerating eﬀect on the synchronous machines. which has a stabilizing eﬀect on the frequency.1 Dynamics of the Generators After a disturbance in the system. is the frequency that can be deﬁned for the so–called centre of inertia (COI) of the system. Generally. 2. Also of interest is the frequency dependency of the load in the system. the exact version of the swing equation will be used 9 . We want to derive a model that is valid for reasonable frequency deviations. deviations from this desired value arise due to imbalances between the instantaneous generation and consumption of electric power. The nominal frequency is assumed to be 50 Hz as in the ENTSOE Continental Europe system (former UCTE).
while the amplitude of ∆ωi is the main concern in frequency control.u.) − Pei (p. In order to convert the torques in eq.u.3) The power can also be expressed in SI–units (e.u.10 50 2.2 49 48.8 0 1 2 t(s) 3 4 5 Figure 2. the predisturbance frequency. Note ˙ that for rotor oscillations the frequency of ∆ωi is often of interest.g. one obtains ∆ωi = ωi . (2.u.1) is formulated for ∆ω.u. (2. The thicker solid curve indicates the average system frequency.) by multiplication with the power base SBi . (2. Also note that for the initial condition ∆ωi (t0 ) = 0 holds if eq.) ω0 is used. 2Hi ωi (2. which yields: ∆ω i = ˙ 2 ω0 (Pmi (p.6 f(Hz) 49.) = T (p.1). Other curves depict the frequency of individual generators. is normally the nominal frequency ωi (t0 ) = ω0 . in MW instead of p. (2. (2. The initial condition for eq.) − Tei (p.1.2) ˙ By deriving eq.)) .u. to describe the dynamic behaviour of generator i: ω0 ωi = ˙ (Tmi (p. 2Hi (2. ωi is the absolute value of the rotor angular frequency of generator i.2) with respect to the time. which represents the rated power of .4 49. the relation ω P (p.1) to power values.8 49.)) . Frequency Dynamics in Electric Power Systems 49.u. Of main interest is usually the angular frequency deviation ∆ωi : ∆ωi = ωi − ω0 .1) with the usual notation. The indices m and e denote mechanical (turbine) and electrical quantities respectively. The frequency in diﬀerent locations in an electric power system after a disturbance.
An illustration of this modelling is depicted in Figure 2. please refer to Appendix A. ωi = ω can be as . ωi (2.2.2. In a highly meshed system. ω0 ωi (2. With further simpliﬁcations. all units can be assumed to be connected to the same bus. (2.2. Furthermore. the generator i.1.4) Note that this is still the exact version of the swing equation. Now. representing the centre of inertia of the system.4) for the n generators in the system yields n 2 i=1 Hi SBi 1 ∆ω i = ˙ ω0 n i=1 ω0 (Pmi − Pei ) . A summation of all the equations (2.3) can be rewritten such that the unit on both sides is MW: 2Hi SBi ω0 ∆ω i = ˙ (Pmi − Pei ) . Dynamic Model of the System Frequency set Pm 11 Pm T G Generator Turbine Pe Pl o a d Load Figure 2. For further details on diﬀerent formulations of the swing equation. eq. they can even be condensed into one single unit. Simpliﬁed representation of a power system consisting of a single generator connected to the same bus as the load.5) Because of the strong coupling of the generation units. which is nonlinear. the goal is to derive the diﬀerential equation for the entire system containing n generators.
11) Eq. ω = ω0 can be assumed for the righthand side. Because of ω = 2πf and ω = 2π f˙ follows ˙ ∆f˙ = f0 (Pm − Pe ) .12) In order to obtain a linear approximation of eq. (2.9) (2.3.14) .11) is illustrated as a block diagram in Figure 2.13) 2HSB The dynamics can also be expressed in terms of frequency instead of angular frequency.11). This yields ω0 ∆ω = ˙ (Pm − Pe ) . (2. (2. Frequency Dynamics in Electric Power Systems – Pe + a */* a/b 2 0 2 HS B 1 s 0 b Figure 2.11). 2HSB (2. (2. Total inertia constant. Block diagram of nonlinear frequency dynamics as in eq. For the frequency ω in the centre of inertia holds as well ω = ω0 + ∆ω . 2HSB ω (2. (2.8) (2. By deﬁning the quantities ω= SB = i i Hi ωi Hi SBi i Hi SBi i SBi Centre of Inertia frequency Total rating. This is a valid assumption for realistic frequency deviations in power systems. Total mechanical power.10) H= Pm = i Pmi Pei i Pe = the principal frequency dynamics of the system can be described by the nonlinear diﬀerential equation ∆ω = ˙ 2 ω0 (Pm − Pe ) .6) (2. Total electrical power.12 Pm 2. (2. sumed for all i.7) (2.3.
1. (2.21) If neither the disturbance nor the oscillations in the transmission system are too large.20) (2. Pm0 = Pe0 and Pm0 = Pload0 + Ploss0 (2.23) ω0 (∆Pm − ∆Pload ) . (2. in the same way as in eq. these approximations are reasonable.2.21). eq. Using eqs.17) are valid. (2. be written as Pe = Pe0 + ∆Pload + ∆Ploss with Pe0 = Pload0 + Ploss0 .13) can now be written as ∆ω = ˙ or equivalently ∆f˙ = f0 (∆Pm − ∆Pload ) . i. The overall goal of our analysis is to derive an expression that gives the variation of ∆ω after a disturbance of the balance between Pm and Pe .4.16) i which can. (2.19) (2.23) can be represented by the block diagram in Figure 2. 2HSB (2.e.15).18) (2. Dynamic Model of the System Frequency 13 A very simple and useful model can be derived if some more assumptions are made. . If the system is in equilibrium prior to the disturbance. (2. Therefore. Furthermore.e. 2HSB (2. the transmission losses after and before the disturbance are assumed to be equal. we deﬁne Pm = i Pmi = Pm0 + ∆Pm . Pe = Pei = Pload + Ploss . (2. (2.15) where Pm0 denotes the mechanical power produced by the generators in steady state and ∆Pm denotes a deviation from that value. The total generated power is consumed by the loads and the transmission system losses.22) Eq. ∆Ploss = 0 . i.15) – (2.
In real power systems.1.24) where f0 • Pload : Load power when f = f0 . dt (2. Linearized model of the power system frequency dynamics. Frequency Dynamics in Electric Power Systems Pload f0 (2 HS B ) s Pm System inertia f Figure 2. This is due to the fact that kinetic energy can be stored in the rotating masses of the motors. A load model that captures both eﬀects is given by f f0 f Pload − Pload = ∆Pload = Kl ∆f + g(∆f˙) (2. • Kl : Frequency dependency. The rotating masses have the following kinetic energy: 1 W (f ) = J(2πf )2 2 (2. This has a stabilizing eﬀect on the system frequency f . a frequency dependency of the aggregated system load is clearly observable. as will be shown in the sequel. 2. which is equal to the power PM consumed by the motor. is given by dW PM = (2.26) dt and ∆PM = d∆W .2 Frequency Dependency of the Loads Loads are either frequencydependent or frequencyindependent.25) The change in the kinetic energy. • g(∆f˙): Function that models the loads with rotating masses. The function g(∆f˙) will now be derived. Apart from a component depending directly on f . large rotating motor loads cause an additional contribution depending on f˙.27) .14 2.4.
2 % per % of frequency variation.29) The values of W0 and Dl are obviously highly dependent on the structure of the load and can be variable over time. the absence of any other control equipment would lead to unacceptable and remaining frequency deviations even for moderate disturbances. . ∂f Dl (2. Dynamic Model of the System Frequency ∆W can be approximated by W (f0 + ∆f ) = 2π 2 J(f0 + ∆f )2 = W0 + ∆W 2 = 2π 2 Jf0 + 2π 2 J2f0 ∆f + 2π 2 J(∆f )2 2W0 W0 = W0 + ∆f + 2 (∆f )2 f0 f0 2W0 ≈ ∆f f0 2W0 d∆f 2W0 ˙ = ≈ ∆f f0 dt f0 15 ⇒ ∆W ⇒ ∆PM (2. As we will see in the next section.5 represents the dynamic load model.6.1. Especially W0 is only a factor in power systems with large industrial consumers running heavy rotating machines. Together with the power system dynamics derived before. . we obtain a dynamical system with a ”proportional/diﬀerential control” caused by the loads. this eﬀect is too small to be able to keep the frequency within reasonable bounds.2. The constant Dl has typical values such that the variation of the load is equal to 0 . The block diagram in Figure 2.28) The frequency dependency of the remaining load can also be written as ∂Pload 1 ∆f = Kl ∆f = ∆f . However. The power system model derived so far is shown in Figure 2. .
Model of power system without control. .5. Block diagram of the dynamic load model. f Pload Frequencydependent loads Pload System load change 1 Dl f Pload 2W0 s f0 Rotating mass loads Pe f0 (2 HS B ) s System inertia f Pm Power generation change Figure 2. Frequency Dynamics in Electric Power Systems 1 Dl f 2W0 s f0 Figure 2.16 2.6.
Parameter H SB f0 DL W0 Value 5 10 50 1 200 Unit s GW Hz Hz MW MW Hz 100 Table 2. it illustrates well the possible frequency rise or decay and the stabilizing selfcontrol eﬀect caused by the frequency dependency of the load. a time plot of the system frequency is shown corresponding to the diﬀerent disturbances. ∆Pload = 500 MW. Dynamic Response of Uncontrolled Power System 17 2. Both loss of generation and loss of load will be shown. However. ∆Pload = 1000 MW (sudden increase of load or loss of generation).1 displays the parameters used in the simulation. Note that this result is purely theoretic as such large frequency deviations could never occur in a real power system because of various protection mechanisms.7. represented by a positive resp. 56 −1000 MW −500 MW −100 MW +100 MW +500 MW +1000 MW 54 System Frequency [Hz] 52 50 48 46 44 0 10 20 30 Time [s] 40 50 60 Figure 2. ∆Pload = −100 MW (sudden loss of load) and ∆Pload = 100 MW.2 Dynamic Response of Uncontrolled Power System Now we will conduct a numerical simulation of the uncontrolled frequency dynamics after a disturbance. In Figure 2. The plot shows the time evolution of the system frequency for a disturbance of (top to bottom) ∆Pload = −1000 MW.2. ∆Pload = −500 MW.2. Table 2.7. . Parameters for time domain simulation of power system. negative step input on the variable ∆Pload .1. Theoretical frequency responses of uncontrolled power system (DL = 1/200 Hz/MW).
In primary frequency control. A non–nominal frequency in the system results in a lower quality of the delivered electrical energy. Furthermore. Since too large frequency deviations in a system are not acceptable. servos. However. in a primarycontrolled power plant. Further. is not discussed here. gates. and deviations from the set values results in a signal that will inﬂuence the valves. like the tripping of a generator. the control task of priority is to bring the frequency back to (short term) acceptable values. etc. is used. the control structures that ensure a constant system frequency of 50 Hz will be described. which in the worst case have to be disconnected. there remains an unavoidable frequency control error because the control law is purely proportional. which has the goal of keeping the frequency during disturbances at an acceptable level. Further explanations follow in chapter 3. This is due to the fact that the utilization of tertiary control reserves is more similar to the electricity production according to generations schedules (dispatch). In comparison with thermal units. . which are based on economic oﬀline optimizations. for example when an area that contains much generation capacity is isolated. Frequency Dynamics in Electric Power Systems 2. automatic frequency control.4 Control Structures for Frequency Control In the following two chapters. This constitutes an even worse stress on the system and can lead to a complete power system collapse. It is also possible that the frequency rises during a disturbance. The control task is shared by all generators participating in the primary frequency control irrespective of the location of the disturbance. The automatic control system consists of two main parts. Many of the devices that are connected to the system work best at nominal frequency.18 2. such that the desired active power output is delivered. The primary control refers to control actions that are done locally (on the power plant level) based on the setpoints for frequency and power. There are at least two reasons against allowing the frequency to deviate too much from its nominal value. The actual values of these can be measured locally. the spontaneous load variations in an electric power system result in a minute–to–minute variation of up to 2%. hydro power plants are more robust and can normally cope with frequencies down to 45 Hz. This alone requires that some form of frequency control must be used in most systems. Tertiary control. which is manually activated in order to release the used primary and secondary control reserves after a disturbance. 2. too low frequencies (lower than ≈ 47 − 48 Hz) lead to damaging vibrations in steam turbines.3 The Importance of a Constant System Frequency In the most common case ∆Pm − ∆Pload is negative after a disturbance. the primary and secondary control.
In many systems.e. also called Load Frequency Control. . a rotating scheme for how the load should be shed. the deployment of tertiary reserves occurs less often. if that is necessary. In the ENTSOE Continental Europe system. power exchanges not according to the scheduled transfers. The secondary control ensures by a special mechanism that this is remedied after a short period of time. Apart from that.8. As the activation of this scheme implies the loss of load in entire regions. Conversely. Note that Load Frequency Control can also be performed manually as in the Nordel power system. Note that primary and secondary control are continuously active also in normal operation of the grid in order to compensate for small ﬂuctuations. The basic control structures described above are depicted in Figure 2. Such a scheme is often called rotating load shedding.9 shows an illustration of the time spans in which these diﬀerent control loops are active after a disturbance.4. the ﬁrst load shedding stage is activated at a frequency of 49 Hz. another undesired eﬀect has to be compensated by secondary control: active power imbalances and primary control actions cause changes in the load ﬂows on the tielines to other areas. i. is devised. which can also be called Automatic Generation Control. This is further discussed in chapter 4. Figure 2. Note that in this control loop the location of the disturbance is considered when the control action is determined: only disturbances within its own control zone (area) are ”seen” by the secondary controller. an automatic scheme is used. Control Structures for Frequency Control 19 In the secondary frequency control.2. causing the shedding of about 15 % of the overall load. it must only be activated if absolutely necessary in order to save the system. the power setpoints of the generators are adjusted in order to compensate for the remaining frequency error after the primary control has acted. Underfrequency load shedding is a form of system protection and acts on timescales well under one second. In the ENTSOE Continental Europe interconnected system.
Temporal structure of control reserve usage after a disturbance (terminology according to ENTSOE for region Continental Europe (former UCTE). . Power Primary Control Secondary Control Tertiary Control Generation Rescheduling 30 s 15 min 60 min Time Figure 2.8. Basic structure of frequency control in electric power systems. Frequency Generator Automatic Generation Control (AGC) identical for synchronous machines in steady state Setpoint Calculation Turbine Governor Valves or Gates Turbine Setpoint from power generation schedule and tertiary control Turbine Controller Speed (Internal turbine control loop) (Primary control loop) Figure 2.20 2. Frequency Dynamics in Electric Power Systems Load Shedding (emergency control) (Secondary control loop) TieLine Powers System Frequency Electrical system Loads Transmission lines Other generators Power.9.
3 Primary Frequency Control In this chapter. where more energy is extracted from the steam. In practice. these turbinegenerator systems can be very large. we will neglect this for the sake of simplicity in the ﬁrst three sections of this chapter. In section 3. These diﬀerent applications will be discussed along with their principal static and dynamic characteristics. diﬀerent steam and hydro turbines and their modelling are discussed and later their eﬀect on the dynamical response will be shown. Instead. which is shown as a block diagram in Figure 3. However. Often the steam is then reheated before it is injected into a medium pressure or low pressure turbine. In the high pressure turbine. The control law. In a big thermal unit of rating 1000 MW.1. is a proportional feedback control. the total amount of necessary primary control reserves are determined by statistical considerations and the distribution on the power plants can vary. As will be shown. a schematic drawing of the primary control is shown in Figure 3. the total length of the turbinegenerator shaft may exceed 50 m. This is also why the control law cannot have an integral component: integrators of diﬀerent power plants could start ”competing” each other for power production shares. It establishes an aﬃne relation between the measured frequency and the power generation of the plant in steady state. there is no coordination between the diﬀerent units. In an interconnected power system. Primary control can also be used in islanded operation of a single generator.1. 3.1 Implementation of Primary Control in the Power Plant For a thermal unit. The turbine governor (depicted here together with the internal turbine controller) acts on a servomotor in order to adjust the valve through which the live steam (coming from the boiler with high pressure and high temperature) ﬂows to the turbines.2. Primary control is implemented on a purely local level. not all generation units need to have primary control equipment. they are implemented entirely on the power plant level. which can lead to an unpredictable and unreasonable distribution of power generation on the available plants. 21 . the mechanisms for primary frequency control are illustrated.4. Note that the turbine dynamics plays in important role in the overall dynamical response of the system. part of the energy of the steam is converted into mechanical energy.
2. LP = Low Pressure turbine. while in practice the measurement can be done on the electrical side.tot T Turbine shaft G f Pe Pmset .22 3. P0 Reference values Figure 3. Schematic drawing of the primary control installed in a thermal unit. . Block diagram describing the primary control law. power Electrical power Electricity grid Boiler Steam G HP Valve Actuation Control signals LP Servomotor Controller/ Governor Measured values f . Pm Internal Turbine Controller Pmset .1. HP = High Pressure Turbine. The measured power value corresponds to Pm in our notation. Primary Frequency Control Reheater Steam Mech.P f 0 .tot Pmset 0 set Pm K 1/ S f f set f0 Proportional control law Figure 3.
3.) S ( p. f ( p. Referring to the block diagram in Figure 3. Static Characteristics of Primary Control 23 3. S = − set. set m0 set Pm .3) Under the assumption that the turbine power controller has an integrating characteristic (ε → 0 when t → ∞ in Figure 3. The speed droop characteristic. f0 and S.u.) P ( p.) f0 ( p.2) (3.) .2.tot + (Pm0 − Pm )=0 . the equation describing the primary control is given by (f0 − f ) · which can be written as S=− or in per unit as set Pm0 1 set set. f ) of the turbine. The position and slope set of the straight line can be ﬁxed by the parameters Pm0 .) Figure 3. Figure 3. Static characteristic of primary control. it follows that in steady set. yields all possible steadyset.tot = − set.2).tot state operating points (Pm .1) f0 − f f − f0 Hz/MW > 0 set.tot ( p.tot 0 set Pm − Pm0 set Pm0 (3.2. We set. In the literature the speed droop characteristics is sometimes also described by ω instead of by f .u. S (3.2 Static Characteristics of Primary Control First. it is of particular interest to study the properties of the primary frequency control in steady state.tot set − Pm Pm − Pm0 f − f0 f .tot state holds Pm = Pm .u.3.tot have chosen to label the horizontal axis with the power Pm which for small deviations of the frequency around the nominal value is identical to the torque T .u.3.u.
and second when the generator is in islanded operation feeding a load. From the speed droop characteristics.5. This is shown in Figure 3.5. P Figure 3. the power produced by the generator can then be determined. Pmset 0 R T Pm G Pe XL f U fG UG fG . not the frequency.2. Figure 3. The turbine governor controls thus only the power. when the generator is part of a large interconnected system. Speed droop characteristics for the case when the generator is connected to an inﬁnite bus (large system).5. it can be modelled with a very good approximation as connected to an inﬁnite bus.4. .tot m ( fG ) Figure 3.4. f f0 fG S f 0 und Pmset are set in the generator 0 P g ( fG ) Pmset . see Figure 3. The third system to be studied is a two machine system. Generator in Large System If a generator is embedded in a large interconnected system.tot P set m0 P set . Primary Frequency Control 3. Generator operating in a large interconnected system. First.1 Role of speed droop depending on type of power system We will now study how the frequency control of a generator will act in three diﬀerent situations. Infinite Bus f 0 .24 3. In steady state the frequency is given by the grid frequency fG (represented by the inﬁnite bus).
In this case the primary control loop will control the frequency.tot m Pe Figure 3. Figure 3. By a voltage controller the voltage U is kept constant and thus also Pe .P Figure 3.3. With the help of the speed droop characteristics of the two systems.tot P set m0 P set . and the speed droop is then the sum of all the individual speed droops of the generators in the two subsystems.6. the generator feeds a load in islanded operation. Pm. The resulting frequency can be determined from the speed droop characteristics. which here is assumed to be a purely resistive load.8. Pmset 0 R T Pm G Pe P e U 2 R R const.9. we will determine how a change in load will be compensated by the two systems.7. Generator in islanded operation. Voltage Control f 0 . Figure 3. Speed droop characteristics for the case when the generator is in islanded operation. what will the changes in Pm.7. Thus. not the power.2 . if we have a change ∆Pload of the overall load.1 .6. U f . In this model the two generators could represent two subsystems. Static Characteristics of Primary Control 25 Islanded Operation As depicted in Figure 3. provides a simple model that is often used to study the interaction between two areas in a large system. f f0 f S f 0 und Pmset are set in the generator 0 f h( P ) Pmset . Two Generator System The two generator system. Figure 3. and f be? .2.
S1 .26 This will be solved in the following way: 3.1 .1 Pmset 0.2 f 0. and thus ∆Pm.2 +∆Pm.1 . S2 ) describe the speed droop characteristics of the two systems g1 and g2 .2 Pload Pm . • From these the sum g3 = g1 + g2 is formed.1 Pm.1 Pload Pmset . S1 0.9.1 P set m .1 . and fN + ∆f be determined. Two generator system. Primary Frequency Control set set • The quantities (Pm0.2 set .2 Pm .2 All these steps are shown in Figure 3.1 G1 Pm .1 +∆Pm.1 .2 Pload Pload Figure 3. f0. S1 ) and (Pm0.2 G2 Pm.1 Pmset . Pm. Speed droop characteristics for a two machine system. S 2 0.1 fN fN f S3 .1 . Pmset . f 0.1 Pm . set set set set • In a similar way: From Pload +∆Pload can Pm0. g 2 Pmset 0.2 .2 .1 m. Pm0. g 3 S2 .8.2 Pmset .2 at f N Figure 3.2 and fN from g3 .2 Pmset . f0.2 P set m . g1 f 0. f 0.1 Pmset . set set • From the given Pload we can determine Pm. ∆P set und ∆f .9.2 . .
3.3. Dynamic Characteristics of Primary Control
27
3.3
3.3.1
Dynamic Characteristics of Primary Control
Dynamic Model of a OneArea System
In this section, we are going to extend the dynamic frequency model introduced in Chapter 2 by the primary control loop in the power plants. For an individual generator, the block diagram has been introduced in the previous section. Following the nomenclature introduced in Figure 3.2, we start with
set,tot set set set Pm = Pm0 + ∆Pm = Pm0 −
1 ∆f , S
(3.4)
set,tot describes the power setpoint of the turbine including the schedwhere Pm set uled value Pm0 and the component imposed by primary control. For the linearized consideration of the power system in ∆ quantities, we set set set Pm0 = ∆Pm0
(3.5)
as the steadystate component cancels out against the other steadystate set,tot translates quantities. The remaining question is how a change in ∆Pm into an actual mechanical power output change ∆Pm . Thus, we regard now the internal turbine control loop as depicted in Figure 3.10.
Pmset Turbine Controller
Turbine
Pmset 0
Kt s
Gt ( s )
Pm
Figure 3.10. Block diagram of the turbine and turbine control dynamics.
From this ﬁgure it follows that ∆Pm (s) = Gt (s) 1 Gt (s) + s Kt
set set ∆Pm0 (s) + ∆Pm (s)
.
(3.6)
If the dynamics of the turbine is neglected (Gt (s) = 1), one obtains ∆Pm (s) = 1 set set ∆Pm0 (s) + ∆Pm (s) 1 + Tt s . (3.7)
Note that the time constant Tt = 1/Kt is fairly small compared with the frequency dynamics of the system. Regarding now only the frequency control set component (∆Pm0 = 0), we obtain in steady state as before 1 ∆Pm = − ∆f . S (3.8)
28
3. Primary Frequency Control
For the case of n controllers for n generators we obtain analogously for controller i 1 ∆Pmi = − ∆f i = 1, . . . , n (3.9) Si 1 ∆Pmi = − ∆f (3.10) Si
i i
where ∆Pm =
i
∆Pmi
(3.11)
is the total change in turbine power. By deﬁning 1 = S we thus have 1 Si (3.12)
i
1 ∆Pm = − ∆f . (3.13) S This can be inserted into the dynamic system frequency model derived in Chapter 2 as depicted in Figure 3.11.
Frequencydependent loads
1 Dl
f Pload
Pload System load change
2W0 s f0
Rotating mass loads
Pe
f0 (2 HS B ) s
Pm
f
1 S
Primary control
P
set m
System inertia
1 1 Tt s
Pmset 0
Turbine dynamics/ control
Figure 3.11. Dynamic frequency model of the power system with primarycontrolled power plants.
3.3. Dynamic Characteristics of Primary Control
29
3.3.2
Dynamic Response of the OneArea System
Now we are going to study the eﬀect of a disturbance in the system derived above. Both loss of generation and loss of load can be simulated by imposing a positive or negative step input on the variable ∆Pload . A change of the set value of the system frequency f0 is not considered as this is not meaningful in real power systems. From the block diagram in Figure 3.11 it is straightforward to derive the set transfer function between ∆Pload and ∆f (∆Pm0 = 0): ∆f (s) = − 1 + sTt ∆Pload (s) (3.14) 1 2W0 2HSB 1 + (1 + sTt ) + ( + )s(1 + sTt ) S Dl f0 f0
The step response for
∆Pload s is given in Figure 3.12. The frequency deviation in steady state is ∆Pload (s) = ∆f∞ = lim (s · ∆f (s)) =
s→0
(3.15)
−∆Pload −∆Pload = = −∆Pload · DR 1 1 1 + S Dl DR 1 1 1 = + DR S Dl
(3.16)
with
(3.17)
In order to calculate an equivalent time constant Teq , Tt is put to 0. This can be done since for realistic systems the turbine controller time constant Tt is much smaller than the time constant of the frequency dynamics TM : Tt ≪ TM = f0 2W0 2HSB ( + ) . SB f 0 f0 (3.18)
This means that the transfer function in eq. (3.14) can be approximated by a ﬁrst order function ∆f (s) = −∆Pload (s) −1 DR ∆Pload = 1 SB SB s + TM s 1 + TM DR s DR f0 f0 ∆f (s) = with Teq = TM DR as the equivalent time constant. 1 SB 1 + TM DR s f0 SB f0 ∆f∞ s (3.19)
or
(3.20)
(3.21)
parameterized as described in the example below) after a step increase in load.12.85 49.9 49. The upper plot shows the system frequency f .04·50 4000 f0 SB Hz/MW = 50 4000 1 2000 Hz/MW 1 80 • Dl = 1 %/1 % ⇒ Dl = • ∆Pload = 400 MW • TM = 10 s Then follows ∆f∞ = −DR · ∆Pload = − and Teq = 10 s · = Hz/MW = Hz/MW 2000 MW Hz 1 · 400M W = −0. f and the frequencydependent load variation ∆Pload . MW + 80 Hz 50 Hz .30 50 System Frequency [Hz] 49.39 s . The lower plot shows the step function in ∆Pload . the increase in turbine power ∆Pm .75 3. Behaviour of the onearea system (no turbine dynamics.23) 2000 MW Hz 1 4000 MW = 0. Primary Frequency Control 0 1 2 3 4 5 6 7 8 9 10 400 Power [MW] 300 200 100 0 0 1 2 3 4 5 Time [s] 6 7 8 9 10 ∆P ∆P load m f ∆ Pload Figure 3.19 Hz (3.8 49.95 49. Example • SB = 4000 MW • f0 = 50 Hz f0 • S = 4% = 0.04 SB = 0.22) + 80 MW Hz (3.
the power exchange PT 12 over the tie line between the areas 1 and 2 has to be modelled. it can be represented by a single bus where all units are connected.3.1 − ϕ0.14. Understanding the interactions between these areas is therefore highly important for the ﬂawless operation of the entire system. to countries. If the power system is highly meshed in this case.1 Pmset 0. a large interconnected power system is always divided into various ”control zones” or ”areas”.13.2 )(∆ϕ1 − ∆ϕ2 ) ∂ϕ1 ∂ϕ2 X (3. P m .24) X where X is the (equivalent) reactance of the tie line.3 Extension to a TwoArea System Up to now we have mostly studied the behaviour of a power system consisting of a single area.27) X By using this model. This is depicted in Figure 3.3. corresponding e. however.2 ) . (3. 2 T2 R2 Pmset 0.1 Load 1 Area 1 Pl o a d .1 G1 T1 R1 PT 1 2 Tieline power flow Pe .3. 2 Load 2 Area 2 f2 Figure 3. For simpliﬁed simulation studies. 2 G2 Pm . . In order to adapt our dynamic frequency model accordingly. This is given by U1 U2 PT 12 = sin(ϕ1 − ϕ2 ) (3.1 Pe . For small deviations (U1 and U2 are constant) one gets ∆PT 12 = or with ∂PT 12 ∂PT 12 U1 U2 ∆ϕ1 + ∆ϕ2 = cos(ϕ0. the block diagram of the power system can be extended as shown in Figure 3. Dynamic Characteristics of Primary Control 31 3.13.2 f1 Pl o a d . a system with two areas can be represented by two single bus systems with a tieline in between them.g.1 − ϕ0. Simpliﬁed representation of a power system with two areas.25) ˆ ∆PT 12 = PT (∆ϕ1 − ∆ϕ2 ) (3. In practice.26) U1 U2 ˆ PT = cos(ϕ0.
1 ) s Pm .32 3. Twoarea dynamic model including tieline ﬂows.1 set Pm 0.1s Turbine dynamics/ control Frequencydependent loads 1 Dl .2 Turbine dynamics/ control Figure 3.2 ˆ 2 PT s PT 12 ˆ 2 PT s PT 12 2W0.14.2 f Pload .2 Pload .1 f0 (2 H1S B .2 f2 1 S2 Primary control P set m .1 s f0 Rotating mass loads Pe.1 PT 12 2W0.1 f1 1 S1 Primary control Pmset .2 System inertia 1 1 Tt s set Pm 0.1 Pload .2 s f0 Rotating mass loads Pe .2 ) s Pm . .2 f0 (2 H 2 S B . Primary Frequency Control Frequencydependent loads 1 Dl .1 f Pload .1 System inertia 1 1 Tt .
The system to be studied is depicted in Figure 3.15.2 SB2 TM. As Area 2 is very big (inﬁnite bus) it follows that TM. We will now study the behaviour after a load change in the smaller area.2 .1 (1 + sTt ) f0 s2 (3.1 SB.1 (s) TM. ∆Pm0. Area 1. Area 1 is much smaller than Area 2.1 .4 Dynamic Response of the TwoArea System For the case regarded here.e. f 0 f1 f2 f0 Figure 3.1 SB1 ≫ ⇒ f2 = constant ⇒ ∆ϕ2 = 0 f0 f0 and consequently ˆ ˆ ∆PT 12 = −∆PT 21 = PT ∆ϕ1 = 2π PT ∆f1 dt (3.2 .2 U2.1 SB.30) ˆT 2π P ∆PT 12 (s) = ∆f1 (s) (3. Dynamic Characteristics of Primary Control 33 3.31) s ˆ −2π PT ∆PT 12 (s) = ∆Pload.1 SB. Model of a two area system.32) ∆f1 (s) = .1 2 1 1 ˆ 2π PT + ( + )s + s Dl.1 (1 + sTt ) f0 (3. Dl .29) (3. f 0 S 2 .3.1 (s): −s ∆Pload. S B .1 (s) TM.3. The bigger one of the two areas can then be regarded as an inﬁnite bus in our analysis.i = 0.W0.15. X 1 Area 2 S1 .1 . it is assumed that one of the areas is much smaller than the other. Dl .W0.2 . S B .1 SB. i.1 1 1 ˆ 2π PT + ( + )s + Dl.1 . 2 TM .3.28) set Without any scheduled generator setpoint changes. the following transfer functions apply for a change in the Area 1 system load ∆Pload.1 TM . Area 1 PT 12 U1 .
1 ∆P T12 0 1 2 3 4 5 Time [s] 6 7 8 9 10 Figure 3. Only primary control is used and the system is parameterized according to Table 3. The upper diagram shows the frequencies f1 in Area 1 and f2 in Area 2.34) and the steady state deviation of the tie line power is ∆PT 12.1 s→0 (3.16.1 ∆ Pm. Figure 3. Primary Frequency Control ∆Pload.85 0 1 2 3 4 5 Time [s] 6 7 8 9 1 2 10 1000 800 Power 600 400 200 0 ∆P load. a new. Step response for the system in Figure 3. This is achieved by increasing the tieline power so the load increase is fully compensated. unscheduled and persisting energy exchange has arisen between the two areas.15. .9 f f 49.1 s is shown if Figure 3.∞ = lim (s · ∆f1 (s)) = 0 s→0 (3.34 The response for 3.1 and the tieline power ∆PT 21 = −∆PT 12 .1 . The lower diagram shows the step in the system load ∆Pload. the turbine power ∆Pm. The steady state frequency deviation is ∆Pload.∞ = lim (s · ∆PT 12 (s)) = −∆Pload.33) (3.1 (s) = ∆f1.95 49.35) The inﬁnite bus brings the frequency deviation ∆f1 back to zero.14 resp.1. While this is beneﬁcial for the system frequency in Area 1. 50 Frequency [Hz] 49.16.
Dynamic Characteristics of Primary Control Parameter H1 H2 SB.1 SB. .2 f0 Dl.2 S1 S2 ˆ PT ∆Pload. Parameters for time domain simulation of the twoarea power system corresponding to Figure 3.1 Dl.1 W0.3.3.2 W0.1.1 Value 5 5 10 10 50 1 200 1 200 35 Unit s s GW TW Hz Hz MW Hz TW MW Hz MW Hz Hz MW Hz MW 0 0 1 5000 1 5000 533.16.33 1000 MW MW Table 3.
3. control valves and governors. However. In a tandem compound unit all sections are on the same shaft with a single generator.1 Turbine Models Steam Turbines Figures 3. which then is converted into electrical energy in the generator. Certain fractions of the total power are extracted in the diﬀerent turbines. Following the nomenclature introduced in Figure 3. the costs are higher and could seldom be motivated. typically TRH = 4 − 11 s.17.18. the linearized dynamic models of the turbines will enter in the block ”Turbine dynamics/control”.6 s. TCH .19. this signal is deﬁned as 1 ctrl set set (∆Pm0 + ∆Pm − ∆Pm ) . gas. and this is modelled by the factors FV HP . If a reheater is installed. is 0.4 Turbine Modelling and Control This section gives an overview of the modelling of both steam and hydro turbines. The power output from the turbine is controlled through the position of the control valves.19 show the most common steam turbines and their models. ∆Pm = (3.36) Tt s The delay between the diﬀerent parts of the steam path is usually modelled by a ﬁrst order ﬁlter as seen in Figures 3. The time constant of the delay between the intermediate and low pressure turbines. In the dynamic frequency model of the power system derived so far (as depicted in Figure 3. or it can be a nuclear reactor. Most modern units are of tandem compound type. In a steam turbine the stored energy of high temperature and high pressure steam is converted into mechanical (rotating) energy. Primary Frequency Control 3. FLP in the models. . and 3. while a cross compound unit consists of two shafts each connected to a generator. The cross compound unit is operated as one unit with one set of controls. FHP .36 3. The aim here is to give an understanding of the basic physical mechanisms behind these models that are very commonly used in simulation packages for the study of power systems dynamics. even if the crossover compound units are more eﬃcient and have higher capacity.10. and 3. 3. The turbine can be either tandem compound or cross compound.17.1 − 0. is in the order of 0. The original source of heat can be a furnace ﬁred by fossil fuel (coal. Their characteristics and behaviour are also brieﬂy discussed. 3.18.4. only a brief qualitative discussion will be provided.11). the time delay is larger.3−0. or oil) or biomass.4 s. Typical values of the time constant of the delay between the control valves and the highpressure turbine. It is outside the scope of these lecture notes to give a detailed derivation and motivation of these models. TCO . which control the ﬂow of steam to the turbines. The valve position is inﬂuenced by the output signal of the turbine controller. FIP .
Single Reheat Steam Turbines Crossover Reheater Valve position Control Valves. . single reheat conﬁgurations. Steam turbine conﬁgurations and approximate linear models. Steam Chest Shaft HP To Condenser Linear Model 1 1 sTCH ctrl Pm Pm Tandem Compound. Nonreheat and tandem compound.17.3. Turbine Modelling and Control 37 Nonreheat Steam Turbines Valve Position Control Valves. Steam Chest HP IP LP LP Shaft To Condenser Linear Model Pm FHP FIP FLP ctrl Pm 1 1 sTCH 1 1 sTRH 1 1 sTCO Figure 3.4.
18. Steam Chest Reheater Valve position VHP HP IP LP LP Shaft To Condenser Linear Model Pm FVHP FHP FIP FHP ctrl Pm 1 1 sTCH 1 1 sTRH 1 1 1 sTRH 2 1 1 sTCO Figure 3. Steam turbine conﬁgurations and approximate linear models. Double Reheat Steam Turbine Crossover Reheater Control Valves. Primary Frequency Control Tandem Compound. . double reheat conﬁguration. Tandem compound.38 3.
3. LP Shaft HP LP LP Valve position Reheater Crossover IP. single reheat conﬁguration.19. Single Reheat Steam Turbine Control Valves. Turbine Modelling and Control 39 Cross Compound. . LP Shaft IP LP LP Linear Model Pm1 FHP FLP 2 ctrl Pm 1 1 sTCH 1 1 sTRH 1 1 sTCO FIP FLP 2 Pm 2 Figure 3. Steam Chest HP. Steam turbine conﬁgurations and approximate linear models.4. Cross compound.
6 0. TRH = 10 s. TCO = 0. single reheat.3 0. It is depicted in Figure 3. For this system the step response is easy to calculate. the conﬁguration with tandem compound.3. .3 0. Step response of system in Figure 3. FLP = 0.21.2 FHP = 0.1 s.21.20.40 3.7 Power from turbine 0.20 can be used.20.4 0.4.1 0 0 5 10 15 20 25 Time (s) 30 35 40 45 50 Figure 3. Simpliﬁed model of tandem compound.9 0. Then a simpliﬁed block diagram according to Figure 3.7 ctrl Pm 1 1 s10 Figure 3. with the following data will be studied: TCH = 0.8 0.5 TRH = 10 s 0.17.7 FIP + FLP = 0. single reheat system in Figure 3.3 0. Pm 0. Figure 3. Primary Frequency Control Step Response To illustrate the dynamics of a steam turbine.3 As TCH ≪ TRH und TCO ≪ TRH . we can assume TCH = TCO = 0 for an approximate analysis.17.3 s FHP = 0. FIP = 0. 1 0.
Schematic drawing of hydro turbine with water paths. since the reservoir is large and the water level does not change during the time scale that is of interest here. to be able to determine the frequency’s dynamic behaviour.3. Bernoulli’s equation for a trajectory between the points P1 and P2 can be written as P2 P1 ∂v 1 2 2 · dr + (v2 − v1 ) + Ω2 − Ω1 + ∂t 2 P2 P1 1 dp = 0 . Hydro Turbines Compared with steam turbines. Figure 3.22 depicts a hydro turbine with penstock and hydro reservoir and deﬁnes the notation that will be used from now on.4. Thus. The power produced by a generator is determined by the turbine governor and the dynamic properties of the turbine. Thus. If the amount of hydro–generated power in a system is not suﬃcient. ρ (3. hydro turbines are easier and cheaper to control. frequency control is primarily done in the hydro power plants if available. Turbine Modelling and Control 41 P1 Length of Penstock = L h = Head Area = A Velocity = v P2 Effective Area= a Output Velocity = v out Figure 3. . the steam turbines have to be included in the frequency control.37) The following assumptions are usually made: • v1 = 0. models for the turbine as well as for the turbine control are necessary.22.
Ω2 − Ω1 = −gh . 2 2 a To get the system into standard form. vout = A v a (3. x = v .39) can be written as dv 1 1 A = gh − v dt L 2L a The maximum available power at the turbine is 1 A3 v 3 1 3 P = ρavout = ρ 2 . • The water is incompressible. • The water pressure is the same at P1 and P2 . Further.37). dt 2 (3. L 2Lu2 (3. (3. determined by the opening of the turbine’s control valve (guide vanes). as L dv 1 2 + vout − gh = 0 .41) (3. (3. is denoted a. p1 = p2 . 2u2 . i.44) x3 y = ρA . i.38) make it possible to write (3.43) are introduced. (Here.e. A y = P . If the penstock’s area is A. x for state. and y for output signal.40) is valid and eq. (3.39) The velocity of the water in the penstock is v. with vout = v2 and the length of the penstock L.42 3. (3. u for control signal.) The system now can be written as x = gh − x2 1 ˙ . a u = .38) The above assumptions together with eq. i.e. ρ does not change with water pressure.e. The eﬀective opening of the penstock. Primary Frequency Control • The water velocity is non–zero only in the penstock. we have used the standard notation.42) 2 . (3.
(3. Block diagram showing model of hydro turbine. ∆u.46) (3. The system corresponding to eq. using eqs. (3. In steady state.45). ˙ u0 L u0 L ∆y = 3y0 2y0 √ ∆x − ∆u . and the state is determined by x0 . can be written as √ ∆x = − 2gh ∆x + 2gh ∆u . To get an idea of the system’s properties.44) can be described with the block diagram in Figure 3. u0 .23. x = 0. (3.3. and small variations around an operating point are studied. u0 u0 2gh 2 3 ∆y = 3ρ Ax0 ∆x − 2ρ Ax0 ∆u . and y0 . Eq.23.4. 2u2 2u3 0 0 (3. which ˙ fulﬁls x0 = u0 2gh . ˙ 2Lu2 2Lu3 0 0 which.44) is nonlinear and a detailed analysis is beyond the scope of these lecture notes.47) . Turbine Modelling and Control u */* x⁄u x 43 x ⁄u 2 2 1 2L  Σ + dx dt x y = P = ρA 2 2u 1 s x y 3 gh L ρA 2 Figure 3. (3. 0 2u2 0 Small deviations ∆x.45) 3 y = ρAx0 . the equations are linearised. and ∆y around the operating point satisfy 2 ∆x = −2x0 1 ∆x + 2x0 ∆u .
Primary Frequency Control t(s) 0 5 10 15 20 ∆y y0 .50) The quantity u0 T = a0 T /A also has dimension of time and is denoted Tw .∆u 1 u0 t(s) y0 .∆u 1 u0 Tw = 5 s Figure 3.48) If eqs. The variation of the produced power. and from the above equations it is apparent that this is the time it takes the water to ﬂow through the penstock if a = A. 1 + su0 T (3. after a step change in the control valve.e. u0 1 + Tw s (3.44 ∆u ∆u 1 3.51) .50) can thus be written as ∆y = y0 1 − 2Tw s · ∆u .47) are Laplace–transformed.24. (3. (3.49) which.47) gives ∆y = y0 1 − 2u0 T s · ∆u . i. That time is denoted T : T = L/ 2gh . (3. The time constant Tw is the time it takes for the water to ﬂow through the penstock when a = a0 or u = u0 . ∆y. leading to ∆x = L/T ∆u .2 . for the operation point where the linearization is done. when inserted in the lower of eqs. Eq. (3. ∆x can be solved from the ﬁrst of the equations. √ The quantity L/ 2gh has dimension of time. u0 1 + u0 T s (3.
0 ctrl Pm 1 TV Pdown Pmin Pmin Pmin . That is evident from the step response to eq.51).25. Figure 3. not all poles and zeros are in the left half plane. The output of the valve can be named then ∆Pm and its input ctrl∗ . ∆Pm Pup ctrl Pm * Pmax 1 s Pmax Pmax . the latter has to be formulated in ∆ quantities as well. Note that. . one zero is in the right half plane. This is usually modelled by a ﬁrstorder element.0 Figure 3. Model of a control valve.2 Steam Turbine Control Valves As stated earlier. taking into account the steadystate valve position. the generated power is increased as a consequence of the increased ﬂow. could be renamed to e. a further dynamics is introduced into the system.g. As the valve cannot be moved at inﬁnite speed.3. if the block diagram is expressed in ∆ quantities.4.51) is of non–minimum phase.24.4. (3. the control input of the steam turbine acts on a valve which inﬂuences the inﬂow of live steam into the turbine. so that the water in the penstock can be accelerated. (3. 3. When the water has been accelerated.e. Note that this block must be inserted between the turbine controller ctrl output ∆Pm and the turbine input when the valve dynamics shall be conctrl sidered. That property of water turbines places certain demands on the design of the control system for the turbines. constraints exist on the ramp rate (speed of the valve motion) and the absolute valve position.25 depicts the corresponding block diagram. depicted in Figure 3. The system has the peculiar property to give a lower power just after the opening of the control valve is increased before the desired increased power generation is reached. i. The physical explanation is the lower pressure appearing after the control valve is opened. Furthermore. Turbine Modelling and Control 45 It is evident that the transfer function in eq. In this case.
the latter of which is of interest here.1. is represented by an integrator with the time constant TG . For steam turbines. A model of a hydro turbine governor is given in Figure 3.03 – 0. Primary Frequency Control 3. Typical values for some parameters of the turbine governor for hydro power.46 3.4.1. Parameter TR TG Tp δ σ Typical Values 2.26.4 s 0.e. Typical values for these parameters are given in Table 3. another type of governor is needed.4. Limits for opening and closing speed as well as for the largest and smallest opening of the control valve are given.06 Table 3.2 – 1 0. this has been discussed in Section 3. which have been shown in Section 3.5 s 0. uopen f umax 1 s 1 1 Tp s 1 TG uclose u u u0 umin TR s 1 TR s Figure 3. the guide vane. The main servo.2. The control servo is here represented simply by a time constant Tp . Because of the particularities of hydro turbines.2. .06 s 0.2 – 0.5 – 7.26.3 Hydro Turbine Governors The task of a turbine governor is to control the turbine power output such that it is equal to a set value which consists of a constant and a frequencydependent part.03 – 0. Model of turbine governor for hydro turbine. i.
the control performance will be unsatisfactory when closedloop stability shall be guaranteed. the total feedback after a frequency change is −(δ + σ). a steam turbine with reheater and a hydro turbine.5 Dynamic Responses including Turbine Dynamics In this section. Note that this is equivalent to the loss of a major generation unit.e.27.54) (3. The transient feedback loop has the ampliﬁcation δ for high frequencies. Increasing the static feedback to a range where a reasonable control performance could be attained will make the system unstable. With static feedback only. e. In Figure 3.52) σ Using eq. σ u0 (3. the transient feedback is zero.53) Thus. the stationary change of power is obtained as ∆P = − 1 P0 ∆f .g. The transient feedback loop provides an additional feedback component during nonstationary operating conditions. (3. is u0i . a transient feedback loop and a static feedback loop. S i = σi P0i and the total speed droop. we will simulate the dynamic responses of a power system consisting of diﬀerent generation unit types with primary control equipment. 3. in the system is given by S −1 = i −1 Si . Thus. Dynamic Responses including Turbine Dynamics 47 The controller has two feedback loops. Si . .55) The transient feedback is needed since the water turbine is a non–minimum phase system as discussed above. In steady state. This feedback decays as steady state is attained. (3. the speed droop is initially lower than its stationary value.3. The eﬀect of three diﬀerent turbine dynamics will be studied: a steam turbine without reheater.3. S. i. (3. the power system response to a load increase disturbance of ∆Pload = +1000 MW is compared. The initial total feedback can be about ten times larger than the static feedback. and the ratio between the frequency deviation and the change in the control valve is given by 1 ∆u = − ∆f . The goal is to improve the understanding of the impact of the turbine dynamics on the overall control behaviour of the system. the turbine dynamics is inserted in the block ”Turbine dynamics/control”. We consider again the onearea power system as presented in Figure 3. As stated before.51). the speed droop for generator i. a nuclear power plant.5.11 with the main parameters given according to Table 3.
The hydro turbine allows the most signiﬁcant frequency decay before the system is brought back to the steadystate frequency.48 Parameter H SB f0 DL W0 TCH TRH TCO S TW TG TR σ δ Value 5 10 50 1 200 3.04 1.3.4 0. It can be seen in the following plots that the frequency control performance is highly dependent on the turbine type. p. s s s p.3 s s s p. Primary Frequency Control Unit s GW Hz Hz MW MW Hz 0 0.04 0. Parameters for time domain simulation of the power system. .u. Table 3. The upper plot shows the frequency response of the uncontrolled power system just as a comparison. the time delays caused by the reheater make the response a lot slower. While steam turbines without reheating equipment do not allow the frequency to decay substantially more than their steadystate deviation.3 8 0.2 6.5 0.5 0.u.u.
Dynamic response of the power system without primary control.5 49 0 5 10 15 20 25 30 Primary control on hydro unit 50 f [Hz] 49 48 0 5 10 15 20 25 30 System load step ∆ Pload ∆ Pload [MW] 1000 500 0 0 5 10 15 Time [s] 20 25 30 Figure 3. . and hydro turbines. steam turbines with reheater.5.27.5 49 0 5 10 15 20 25 30 Primary control on unit with reheating 50 f [Hz] 49.3. Dynamic Responses including Turbine Dynamics Uncontrolled power system 50 f [Hz] 48 46 0 5 10 15 20 25 49 30 Primary control on unit without reheating 50 f [Hz] 49. with primary control on steam turbines without reheater.
75 0 1 2 3 4 5 6 7 8 9 10 400 Power [MW] 300 200 100 0 0 1 2 3 4 5 Time [s] 6 7 8 9 10 ∆ Pload ∆ Pm f ∆ Pload Figure 3. and the frequencydependent load f variation ∆Pload .3 s). .50 3.95 49. one more illustration of the eﬀect of the turbine dynamics shall be given. 50 System Frequency [Hz] 49. The upper plot shows the system frequency f . The result is shown in Figure 3.28.28.8 49.12 is now simulated again with a steam turbine model without reheater (TCH = 0.12) after a step increase in load. Behaviour of the onearea system (including dynamics of steam turbine without reheater. otherwise same parameterization as in Figure 3. For a direct comparison of a primarycontrolled onearea system without and with turbine dynamics. the increase in turbine power ∆Pm .9 49. the exemplary system simulated in Figure 3. Primary Frequency Control Finally. The lower plot shows the step function in ∆Pload .85 49.
4. additional control actions must be taken. both during steady state and dynamic conditions.1 Static Characteristics of AGC The overall purpose of the Automatic Generation Control comprises two main aspects: • Keep the frequency in the interconnected power system close to the nominal value. countries. in an interconnected system. The mechanisms of AGC that enable it to fulﬁll these requirements will be outlined in the sequel. The LFC can be done either manually through operator interaction or automatically. the role of the primary frequency control was dealt with.1 • Restore the scheduled interchanges between diﬀerent areas. The characteristics of AGC will be studied in the subsequent sections. This is done through the LoadFrequency Control (LFC). To restore the frequency and the scheduled power interchanges. It was shown that after a disturbance a static frequency error will persist unless additional control actions are taken. deviations of up to ±0. the primary frequency control might also change the scheduled interchanges between diﬀerent areas in an interconnected system.g. Simple models that enable the simulation of the dynamic behaviour during the action of frequency controllers will also be derived and studied. In the previous chapter. North America. In many systems. or loadfrequency. control of power systems will be discussed.1 Hz from the nominal value (50 or 60 Hz) is deemed as acceptable in steady state. even tighter tolerance bands are applied. e. Furthermore. 1 51 . In the latter case it is often called Automatic Generation Control (AGC). In some systems.4 Load Frequency Control In this chapter the secondary. while recently in UK the tolerance band has been relaxed somewhat.
Load Frequency Control Area 2 G PT 1 PT 2 G G set PAGC 2 set PAGC1 G G G f1 AGC1 f2 AGC2 PT 1 Tie line power for Area 1 = j 1 PT j1 Sum over all tie lines Figure 4. A common way is to implement this as a proportionalintegral (PI) controller: ∆PAGCi = −(Cpi + 1 )∆ei sTN i (4. A block diagram of such a controller is given in Figure 4. . N . will adjust the power reference values of the generators participating in the AGC. AGC1 and AGC2 .1) where Cpi = 0.2) Deﬁning now ∆PT i = j∈Ωi j j (PT i − PT 0i ) . (4.1. there are N controllers AGCi .1 . .2. Two area system with AGC. .3) the ACE can be written as ACEi = ∆PT i + Bi ∆f i = 1.4) The set Ωi consist of all areas connected to area i for which the tie line j powers should be controlled to the set value PT 0i . ACEi for area i. (4. . The two secondary frequency controllers. The constants Bi are . The ACEs are in this case: ACEi = j∈Ωi j j (PT i − PT 0i ) + Bi (f − f0 ) i = 1. . .52 Area 1 G 4. Consider a two area system as depicted in Figure 4. . . The error ∆ei is called Area Control Error. . 2. N . . one for each area i. (4. . We will now consider a system with N areas. 1. . 200 s. In an N area system.0 and TN i = 30 . 2.1.
i. f is identical for all areas. . In steady state. The goal is to bring all ACEi → 0. in total N + 1 variables. Selection of Frequency Bias Factors Consider the two area system in Figure 4. Before the AGC has reacted we have a frequency deviation of ∆f in both areas.5) and consequently a solution can be achieved. the generation must be increased in area 2 with ∆Pload after the AGC has reacted. f0i = f0 for all i.e. and f is also the same in steady state for all areas. f0 = f . Since we have N equations (ACEi = 0). i (4. we need one more equation. As a ﬁfth equation we have the power balance: ∆PT i = 0 . i. The variables are thus ∆PT i (N variables) and f .4.1.2. then the system will j j settle down to an operating point where PT 0i = PT i for all tie line powers. The load is now increased with ∆Pload in area 2. Static Characteristics of AGC 53 PT 0i PTi ei AGCi PAGCi Bi f f f0 Figure 4.1. The time constants of the AGC is chosen such that it reacts much slower than the primary frequency control. and we assume that the frequency is controlled back to the reference value. If the sum j of the reference values of the tie line powers PT 0i is 0. Control structure for AGC (∆ei = Error = ACEi = Area Control Error for area i) called frequency bias factors [MW/Hz]. i. It is assumed that the frequency references are the same in all areas. If the tie line power should be kept the same.e.e.
2 Dynamic Characteristics of AGC As in the previous chapter on primary frequency control. there is obviously no tieline power to be controlled. (4. the AGC only fulﬁlls the purpose of restoring the nominal system frequency. The two ACEs can now be written as ACE1 = ∆PT 1 + B1 ∆f = ∆PT 1 + B1 (−S1 ∆PT 1 ) = ∆PT 1 (1 − B1 S1 ) (4. The interested reader is welcome to calculate the transfer function between the load disturbance and the frequency as an exercise. If we set B1 = 1/S1 we see from eq. In this case.1) has an integrating part. and B = 1/S.4 shows a plot of the system frequency which is brought back to the nominal value.2.3 shows the known block diagram which has been extended by the proportionalintegral control law of the AGC.1 Onearea system In a singlearea power system. If B2 = 1/S2 is chosen the ACE in area 2 becomes ACE2 = −∆Pload (4. As the way of extracting inputoutput transfer functions from block diagrams has already been presented in the last chapter. TN = 120 s.9) ∆f = −S2 (∆Pload − ∆PT 1 ) In this case it is desirable that the AGC controller in area 1 does not react. Figure 4. Load Frequency Control (4. This is called Non Interactive Control. 4. In a multiarea case this corresponds to selecting Bi = 1/Si for all areas.8) that ACE1 = 0.10) This means that only controller 2 reacts and the load increase ∆Pload is compensated for in area 2 by the PI control law as stated in eq.8) and ACE2 = ∆PT 2 + B2 ∆f = ∆PT 2 + B2 (−S2 (∆Pload − ∆PT 1 )) = ∆PT 2 (1 − B2 S2 ) − B2 S2 ∆Pload (4.1). we omit the analytical derivation here for shortness. as long as the controller in eq. The choice according to Non Interactive Control has been found to give the best dynamic performance through a number of investigations. The parameters of the system are as presented in Table 3. . Figure 4. (4. we will now develop a dynamic model of the power system where the newly introduced AGC is included. 4. all positive values of Bi will guarantee that all ACEi → 0.54 which means that for area 1 ∆f = −S1 ∆PT 1 and for area 2 4. Cp = 0.6) (4.3 (turbine dynamics neglected). (4.17. However.7) and ∆PT 1 = −∆PT 2 .
4.3.4.05 50 49.2.85 49.95 f [Hz] 49. 50.7 with ACG without AGC 0 100 200 300 400 500 Time [s] 600 700 800 900 Figure 4.1 50. Dynamic response of the onearea system equipped with primary control and AGC compared with the same system without AGC. Dynamic Characteristics of AGC Frequencydependent loads 55 1 Dl f Pload Pload System load change 2W0 s f0 Rotating mass loads Pe f0 (2 HS B ) s Pm f 1 S Primary control P set m System inertia 1 1 Tt s P set m0 Turbine dynamics/ control Psched PAGC (C p 1 ) s TN ACE system) (one area B AGC frequency bias Figure 4. .9 49.75 49. Dynamic model of onearea system with AGC.8 49.
Because of the secondary control (AGC) one obtains the step response in Figure 4.95 49. which initially compensates the disturbance and then slowly decays to 0.2 ∆P T21 Figure 4. The lower diagram shows the control action of the AGCs in Area 1 ∆PAGC. The upper diagram shows the system frequencies of Area 1 and Area 2.6 with AGC.2 which is largely uninﬂuenced by the disturbance.1 = 1000 MW in Area 1 is in this case fully compensated by the generators in Area 1.5.9 49.2 Twoarea system – unequal sizes – disturbance response Here we consider the same twoarea power system as in section 3.05 System frequency [Hz] 50 49. the tie line power ∆PT 21 is shown. . The corresponding block diagram including the AGC is shown in Figure 4.1 which compensates the power deﬁcit and in Area 2 ∆PAGC.2.1 ∆ PAGC.3.56 4.8 f1 f2 0 100 200 300 400 500 600 1000 800 600 P [MW] ∆P 400 200 0 −200 0 100 200 300 Time [s] 400 500 600 AGC.6.4 where Area 1 is assumed to be much smaller than Area 2. Step response for the system in Figure 4. The load increase of ∆Pload. Load Frequency Control 4. Furthermore.85 49. 50.5.
Dynamic Characteristics of AGC 57 1 Dl .2. Dynamic model of twoarea system with AGC.2 (C p 2 1 ) s TN 2 Figure 4.1 1 1 Tt s (C p1 1 ) s TN 1 PT 12 PT 12 ˆ 2 PT s 1 1 Tt s B2 Psched .1 Pload .4. .2 PAGC .6.
Saturation of the control variables (due to limited amounts of contracted control reserves) is not modelled. The reason for this timing is the trading on electricity markets which is usually done in an hourly framework. which means that the probability of the signal to assume a certain value is dependent on the previously assumed signal values. Another important impact on the AGC arises from schedule changes that occur at the full hour. we only simulated the dynamic response of the power system to stepwise disturbances in the load. the dynamic model of a steam turbine without reheater is included.7 demonstrates the response of the twoarea system to a stochastically varying load.) in the control area. . the primary controls are equipped with a dead band of ±10 mHz.2 are shown in the plot. To increase the degree of realism. These normaloperation imbalances are due to load ﬂuctuations. as well as just after the full hour (then in the opposite direction) arise.58 4. the automatic control loops of primary and secondary control are always active to compensate for the continuously arising small imbalances between generation and load. This eﬀect. During the everyday normal operation of a power system. Thus.2. is disregarded in the present setup. only the input time series ∆Pload. and. The system is parameterized again according to Table 3. As often found in practice. In this context. imbalances just before the full hour (either positive or negative). unequal sizes – normal control operation So far. however. however.3 Twoarea system. the diﬀerence between ﬂuctuation and prediction errors is constituted by the time scales on which they occur: while the term ”ﬂuctuation” is usually deﬁned as largely uncorrelated stochastic noise on a second or minute time scale. The latter usually shows a certain autocorrelation. Figure 4. in the presence of intermittent infeeds (from wind power etc. infeed ﬂuctuations and infeed prediction errors. ”prediction error” usually refers to a mismatch of predicted and actual consumption/infeed on a 15minute time scale. load prediction errors. Load Frequency Control 4. Because of ramping actions of power plants which often deviate from the ramping proﬁles they are supposed to follow.1 and ∆Pload. A detailed discussion of the synthesis of the load ﬂuctuation and prediction error signals is beyond the scope of this lecture.1.
4.2. Dynamic Characteristics of AGC
59
System frequencies System frequency [Hz] 50.15 50.1 50.05 50 49.95 0 10 20 30 Activity of secondary control 600 400 P [MW] 200 0 −200 −400 0 10 20 30 Activity of primary control 400 200 P [MW] 0 −200 −400 ∆ Pset m,1 ∆ Pset /1000 m,2 0 10 20 30 40 50 60 40 50 ∆ PAGC,1 ∆ PAGC,2 /1000 ∆ PT21 60 40 50 f1 f2 60
Stochastic load fluctuation and prediction error 400 200 P [MW] 0 −200 −400 ∆ Pload,1 ∆ Pload,2 / 1000 0 10 20 30 Time [min] 40 50 60
Figure 4.7. Normal operation of the two area system. The usual small mismatches between generation and demand are modelled by a stochastic disturbance in the load.
60
4. Load Frequency Control
4.2.4
Twoarea system – equal sizes, including saturations – disturbance response
Finally we brieﬂy investigate the dynamic behaviour of a twoarea system where both areas are of the same size. Turbine dynamics are again neglected in this scenario. Here, saturations of the control inputs which represent the limited amount of contracted primary and secondary control reserves are modelled (primary control: ±75 MW, secondary control: ±350 MW). A graphical representation of the saturations in a block diagram is omitted here for shortness. Referring again to Figure 4.6, the saturation blocks set would have to be inserted in the signal path of ∆Pm (primary control) and ∆PAGC (secondary control). Furthermore, countermeasures against integrator windup have to be taken, which shall not be discussed here. Figure 4.8 shows the step response (∆Pload,1 = 400 MW) for the system with two equally sized areas. It is clearly visible that the AGC of Area 1 cannot compensate for the deviation and remains at its saturation of +350 MW. The system frequency is not brought back to 50 Hz. In real power system operation, the ancillary service dispatcher in the control center would now manually activate tertiary control reserves in order to relieve the saturated secondary reserves.
50 49.98 49.96 f [Hz] 49.94 49.92 49.9 49.88 0 100 200 300 400 500 600 ∆ f1 ∆ f2
400 300 P [MW] 200 100 0 −100 ∆ PT12 ∆ PAGC,1 ∆ PAGC,2
0
100
200
300 Time [s]
400
500
600
Figure 4.8. Step response for the system in Figure 4.6 with AGC, equal areas with a rated power of 10 GW each. Because of the saturated secondary control, the power system frequency cannot be brought back to the nominal value of 50 Hz.
5
Synchronous Machine Model
Almost all energy consumed by various loads in an electric power system is produced by synchronous machines, or, more correctly, the conversion from the primary energy sources, like water energy, nuclear energy, or chemical energy, to electrical energy is done in synchronous machines with a mechanical intermediate link, the turbine. This is true in larger power systems, but not always in smaller systems like isolated islands, power supply of equipment in deserts, or other smaller systems. In these systems, the energy can come from asynchronous generators, for example in wind generation units, batteries, or some other source of electrical energy. In systems with synchronous generators, these have an extremely important part in many dynamic phenomena. Thus, it is very important to develop usable and realistic models of the synchronous machines. In the previous chapters, mainly the mechanical properties of the synchronous machines have been modelled using the swing equation, while a very simplistic model of the electrical properties of the synchronous machine has been used. In this chapter, a more general, detailed model of the electric parts of the synchronous machine will be derived. The simple models used earlier will be justiﬁed. It should be emphasized that the description here aims towards the development of models usable for studying dynamic phenomena in the power system. It is not the purpose of these models to give a detailed and deep understanding of the physical functions of the synchronous machine. Of course, it is desirable to have a good insight into the physics of the synchronous machine to be able to derive appropriate models. For a detailed discussion of these aspects, books and courses dealing with the theory of electrical machines should be studied.
5.1
Park’s Transformation
Park’s transformation is a phase transformation (coordinate transformation) between the three physical phases in a three phase system and three new phases, or coordinates, that are convenient for the analysis of synchronous machines. This transformation is also known as the dq–transformation or Blondel’s transformation. A reason why the transformation is suitable can be derived from Figure 5.1. It is obvious that the phase quantities in the a–, b–, and c–phases will vary periodically in steady state. Further, the self and mutual inductances between stator circuits and rotor circuits will vary with the rotor position. Instead of performing all computations in the ﬁxed stator system, the stator 61
be applied to machines with an arbitrary number of pole pairs. currents. To make the system complete. and one in an axis perpendicular to this. the denominations d–axis and q–axis will be used.1.1 is a simpliﬁed picture of a synchronous machine and should only be viewed as an intuitive basis for the transformation given below. a third component corresponding to the zero sequence must be deﬁned.1 has one pole pair. and the other is the quadrature axis (q–axis). as a consequence of the reasoning above. time dependent. and the connection between the phase currents and the trans . From now on. Synchronous Machine Model aaxis daxis θ ib fb n′ iQ iD Direction of Rotation sa iD baxis qaxis Figure 5. two orthogonal axes are deﬁned as shown in Figure 5.1: One along the axis in which the current in the rotor windings generates a ﬂux. and ﬂuxes can be transformed to a system that rotates with the rotor. sc iF iF iQ n′ fa ia caxis n′ fc ic sb quantities voltages. but Park’s transformation can.62 5. Park’s transformation is. Deﬁnition of quantities in Park’s transformation. of course. The ﬁrst is the direct axis (d–axis). Thus. Figure 5. The machine in Figure 5.
b–. xc )T x0dq = (x0 . It should be pointed out that ia .2) is obtained and the time dependence in the transformation is obvious.6) A mnemonic for Park’s transformation can be obtained from Figure 5.1 by projecting the a–. xabc = (xa .5) The inverse transformation is then given by (5. θ = ωt + θ0 (5. 3 2π 2 ia cos θ + ib cos θ − 3 3 2 2π −ia sin θ − ib sin θ − 3 3 + ic cos θ + 2π 3 2π 3 . Now.3) can be deﬁned. Park’s Transformation formed i0 i d i q 63 currents is given by = = = 1 (ia + ib + ic ) . and it can easily be shown that P −1 = P T . like voltage.7) (5. xd . ib . Equation (5.5. (5.1) − ic sin θ + If the a–axis is chosen as reference. This is reﬂected in the expression for the momentary power that is produced in the stator windings p = ua ia + ub ib + uc ic = uT iabc = abc (P −1 u0dq )T P −1 i0dq = (P T u0dq )T P −1 i0dq = uT P P −1 i0dq = uT i0dq = 0dq 0dq u0 i0 + ud id + uq iq .1. (5. With this notation. x can here be an arbitrary quantity. xb . and c–axes onto the d– and q–axes in the ﬁgure. (5. . Park’s transformation can be written as x0dq = P xabc with P = √ √ √ 1/ 2 1/ 2 1/ 2 2 cos θ cos(θ − 2π ) cos(θ + 2π ) . 3 3 3 2π − sin θ − sin(θ − 3 ) − sin(θ + 2π ) 3 xabc = P −1 x0dq .8) . current. xq )T (5. and ic are the real physical phase currents as functions of time and not a phasor representation of those.4) (5.7) implies that Park’s transformation is an orthonormal transformation. or ﬂux.
Equations (5. it is thus important to make sure that the deﬁnition of Park’s transformation used is the same as one’s own. which is lagging by 90◦ .9) shows that the introduced transformation is power invariant. cf.64 5.12) i. Figure 5. E. pure DCquantities (time independent) in the dq–system with the zero sequence component equal zero.8) can therefore be written as p = ua ia + ub ib + uc ic = u0 i0 + ud id + uq iq .7) have been used.10) It is particularly of interest to study how zero sequence. and positive sequence quantities are transformed by Park’s transformation. The rotor windings produce a ﬂux linkage that mainly lies in the direction of the d–axis. (Show this!) .11) 3 2π sin(θ + α + 3 ) is transformed to 0 3x sin(α) . When using equations from some book or paper. the d– and q–axes have moved by an angle ωt with the angular speed of the rotor ω. The rotor’s d–axis will hence be in position θ = ωt + δ + π . The zero sequence component vanishes also in this case. Synchronous Machine Model Here. It is comparatively easy to show that a pure zero sequence quantity only leads to a contribution in x0 with xd = xq = 0.6) and (5. the phasor for E leads by an angle δ before the phasor for the terminal voltage U .7). It should be pointed out that there are several diﬀerent variants of Park’s transformation appearing in literature. which is a consequence of Equation (5. They can diﬀer from the form presented here by the direction of the q–axis and by constants in the transformation matrix. That ﬂux induces an electromagnetic ﬁeld. wrong results might be obtained.9) Equation (5. 2 (5. Equation (5. the negative q–axis thus leads by an angle δ before the phasor for the voltage along the a–axis.1. hence in the direction of the negative q–axis. (5.e. − cos(α) x0dq (+) = √ (5. A pure negative sequence quantity gives rise to quantities in d– and q–directions that vary with the angular frequency 2ω. A pure positive sequence quantity sin(θ + α) √ xabc (+) = 2x sin(θ + α − 2π ) (5. negative sequence. For generator operation. Otherwise. At t = 0. For t > 0.
(5. The ﬂux linkages in the stator windings (Ψabc ) and in the windings F. aQ Qa Q L = L = −M sin(θ − 2π ) .. and Q (ΨF DQ ) depend on the currents in these windings according to Ψabc ΨF DQ = Labc.F DQ: LF F = LF . the matrix elements are calculated.F DQ iabc iF DQ . Q. (5. The following approximation for inductances L(θ) proves to be useful: L(θ) = Ls + Lm cos 2θ with Ls > Lm > 0.14) Lab = Lba = −Ms − Lm cos(2θ + π ) .abc : L = L = M cos(θ − 2π ) . . These will depend on the rotor position. LF DQ.F DQ and LaF = LF a = MF cos θ . 3 • Labc. . 3 (5. respectively.F DQ LF DQ. Keeping that in mind. .F DQ are 3 × 3 matrices with self and mutual inductances as matrix elements. Quantities related to these windings are denoted with the indices D. LDQ = LQD = 0 . • Labc.2. and F. hence they are time dependent. The Inductance Matrices of the Synchronous Machine 65 5. 3 L = L = −M sin θ . The same holds for the machine constant Ms that will be used in the case of mutual inductances.13) Labc. cQ Qc Q 3 • LF DQ. (5. bQ Q Qb 3 L = L = −M sin(θ + 2π ) . a ﬁeld winding is considered. 3 Lbc = Lcb = −Ms − Lm cos(2θ + π) .5. cF Fc F 3 LaD = LDa = MD cos θ .16) LF D = LDF = MR .abc : Laa = Ls + Lm cos 2θ . LF Q = LQF = 0 . where Ls and Lm are machine constants that can be identiﬁed from measurements. L = LDb = MD cos(θ − 2π ) . bb s m 3 Lcc = Ls + Lm cos(2θ + 4π ) .abc .abc LF DQ. .abc Labc. Lac = Lca = −Ms − Lm cos(2θ + 5π ) . bF F Fb 3 L = L = M cos(θ + 2π ) .2 The Inductance Matrices of the Synchronous Machine In the following. LQQ = LQ . LF DQ. L DD = LD . D. a synchronous machine with one damper winding in d– and one in q–axis and. of course.15) 3 bD LcD = LDc = MD cos(θ + 2π ) . L = L + L cos(2θ − 4π ) .
0 kMQ 3 2 (5. They can be called exact in an ideal machine.21) and 0 0 kMD 0 . I (5.F DQ LF DQ. LF DQ. (5. It is now natural to transform the abc–components in Equation (5. an extended transformation given by Pex = P 0 0 I .18) with the inductance matrix given by L0dq.5) and a 3 × 3 unit matrix I is used. where the inductance matrix in (5.24) (5. not quite exact.14) – (5.F DQ i0dq iF DQ . The resulting inductances are of course.17) with P according to (5.F DQ LF DQ.66 5.abc LF DQ.16) are constants and depend on the design of the synchronous machine.19) The virtue of the Park’s transformation is apparent in the following equation.13) to 0dq–components.F DQ LF DQ. For a real synchronous machine. of course.23) . L = Ls + Ms + 3 Lm .0dq = 0 Ld 0 . (5. It should be emphasized that the model developed here is for use in computations where the synchronous machines are part of a larger system. (5.F DQ P −1 0 0 . Synchronous Machine Model All inductances with only one index in Equations (5. where spatial harmonics and other unsymmetries are neglected.22) (5. For this. the approximations are usually very good and lead to fully acceptable results for the computations and analyses treated here.0dq LF DQ. as mentioned before.0dq = LT 0dq.20) L0dq.0dq L0dq.F DQ where 0 = kMF 0 k= and.18) is computed L0 0 0 (5.0dq L0dq.abc Labc. L0dq. The model is not primarily aimed at studies of the internal quantities in the generator. 0 0 Lq with L0 = Ls − 2Ms .F DQ = P 0 0 I Labc.0dq LF DQ. The result is Ψ0dq ΨF DQ = L0dq.F DQ . 2 d 3 Lq = Ls + Ms − 2 Lm .
(5. uc = −rc ic − Ψ .F DQ has. 5.and qdirections are decoupled. • The quantities in d.3 Voltage Equations for the Synchronous Machine For the three stator circuits and the three rotor circuits the following relations can be written: ˙ ua = −ra ia − Ψa . 0 0 LQ Two important observations can be made from Equations (5.F DQ = MR LD (5. Voltage Equations for the Synchronous Machine 67 daxis d ω D F qaxis q Q Figure 5. (The induction matrix is block diagonal: one 2 × 2 matrix and one 1 × 1 matrix.20)–(5. ˙b .3. Schematic picture of the transformed system.25) 0 .) LF DQ.2.26) u = −rb ib − Ψ b ˙c .5.) The second observation above leads to a picture of the transformed system according to Figure 5. but for completeness it is repeated here.25): • The inductances in the inductance matrix in Equation (3.2. not changed. of course.18) are not dependent on time. (One damper winding in the d–axis and one in the q–axis. LF MR 0 LF DQ.
which in most cases is true. If Equation (5. (5.27) The vector uF DQ is deﬁned as uF DQ = (−uF .29) while the other vectors are deﬁned as before. Rabc and RF DQ are diagonal 3 × 3 matrices. and if ra = rb = rc = r. Sign convention for stator and rotor circuits.28) is multiplied by Pex according to Equation (5.26) and (5. u0dq uF DQ =− P Rabc P −1 0 0 RF DQ i0dq iF DQ − ˙ P Ψabc ˙ ΨF DQ . 0)T .27) can be written more compactly in vector form.31) is valid. (5.28) ˙ uF = rF iF + ΨF . all quantities are transformed to the dq–system. and Equations (5. 0 = rQ iQ + Ψ (5. 0.30) The matrix P Rabc P −1 is denoted R0dq .3.68 5. R0dq = Rabc = rI (5. ˙ 0 = rD iD + ΨD . uabc uF DQ =− Rabc 0 0 RF DQ iabc iF DQ − ˙ Ψabc ˙ ΨF DQ .17). Synchronous Machine Model Figure 5. . i. (5.e. ˙Q .
To further simplify the expressions that were obtained. Equation (5. only one damping winding in the q–axis was considered. = LAQ . uq = −riq − Ψ and ˙ −uF = −rF iF − ΨF . which leads to d ˙ ˙ ˙ Ψ0dq = (P Ψabc ) = P Ψabc + P Ψabc . dt and thus ˙ ˙ ˙ ˙ ˙ P Ψabc = Ψ0dq − P Ψabc = Ψ0dq − P P −1 Ψ0dq . expressions for the dependencies of the ﬂux linkages on the currents in the diﬀerent windings were derived. also the last term on the right hand side must be expressed in these.) We introduce 3 2 MF 3 2 MQ = 3 2 MD = MR = LAD . (5. it is important to remember that ˙ P =0 .33) (5. ˙ (5. (5.34) (5. 0 = −rQ iQ − Ψ (5. the per unit system for the diﬀerent windings is now introduced so that all mutual inductances in the d–axis are equal.30) can hence be written as u0dq uF DQ =− R0dq 0 0 RF DQ i0dq iF DQ − ˙ Ψ0dq ˙ ΨF DQ ˙ P P −1 Ψ0dq 0 (5. d ˙ q − ωΨd . Voltage Equations for the Synchronous Machine 69 To get Equation (5.32) ˙ Some trivial computations show that the matrix P P −1 ˙ P P −1 0 0 0 = 0 0 ω . but in a more general case several damper windings can be considered. (In our case.30) expressed solely in dq–quantities.5.3. ˙ 0 = −rD iD − ΨD . and all in the q–axis are equal.36) The voltage equations in the dq–system can thus be written in component form as ˙ u0 = −ri0 − Ψ0 . Since P is time dependent. ˙Q .38) In the previous section.39) .35) can be expressed as + .37) u = −rid − Ψd + ωΨq . 0 −ω 0 (5.
which justiﬁes that the ﬁrst ones are often neglected. Ψ = Ld id + LAD (iF + iD ) . during which diﬀerent couplings between the circuits in Figure 5. 5. In Equation (5.4.40) (5. Synchronous Machine Model The diﬀerent ﬂuxes can then be written as Ψ0 = L0 i0 . This makes it possible to deﬁne new inductances describing the synchronous machine during these time periods. In Figure 5.37). It can be shown that the terms Ψd and Ψq are in most applications much smaller than ωΨd and ωΨq . In steady state. These time constants indicate also the duration of the time intervals. These equations together with a description of the external system unequivocally determine the behaviour of the synchronous machine during diﬀerent disturbances.70 5. and Subtransient Inductances The complete description of the synchronous machine given in the previous section can be simpliﬁed and made more understandable from a physical standpoint if a number of new parameters are introduced. we observe that the emf in d– and q–direction consists of two terms: one that is a time derivative of the absolute value of the ﬂux linkage and one that arises because the ﬁeld winding is rotating.40) and (5. and these inductances will be derived in this section. the time constants that specify how fast the currents in the damper windings decay are derived in the next section. the induced currents in the damper windings and in the ﬁeld winding vanish. The ﬁrst of these is usually called stator transient and the other rotational emf.4 prevail.41) now describe the electrical dynamics of a synchronous machine completely. Earlier.4 Synchronous. Ψ = LD iD + LAD (id + iF ) . D ΨQ = LQ iQ + LAQ iq . the ﬁrst of these vanishes.38) together with Equations (5.37) and (5. Further. They are . These new parameters can be derived from the already deﬁned ones. d Ψq = Lq iq + LAQ iQ . (5. a graphical description of these equations is given.41) Equations (5. Transient. After a disturbance diﬀerent time intervals can be distinguished. and ΨF = LF iF + LAD (id + iD ) . during which the derived inductances determine the behaviour of the synchronous machine. In steady state or a suﬃciently long time after a disturbance. the synchronous inductances Ld and Lq were deﬁned. and the whole emf is created by the ˙ ˙ rotation of the ﬁeld winding.
repeated here for completeness.4 it is thus obvious that the inductances seen from the generator terminals are in . In steady state the currents iD and iQ = 0. Lm > 0 and thus Ld > Lq . Graphical description of the voltage equations and the coupling between the equivalent circuits. so there is no inﬂuence from the damper windings on the stator circuits.4. while Lm ≈ 0 for machines with round rotors leading to Ld ≈ Lq .5.4. Ld = Ls + Ms + 3 Lm 2 Lq = Ls + Ms − 3 Lm 2 (5.42) For a synchronous machine with salient poles. and Subtransient Inductances 71 rF + uF iF rD LD iD rQ LQ iQ L0 L AQ Lq ωψ d + r0 i0 + u0 L AD Ld ωψ q + rq iq + uq LF rd id + ud Rotor Circuits Stator Circuits Figure 5. Since in steady state the rotor current iF is a pure dc current the only inﬂuence from the ﬁeld current on the stator circuits is through the rotational voltages ωΨd and ωΨq . Synchronous. From Figure 5. Transient. These inductances describe the synchronous machine in steady state as can be seen from Figure 5.4. like a hydro power generation unit.
(5.45) is superimposed on the terminal voltages. e. It is assumed that the machine is in steady state before the disturbance and a voltage change of the form ∆ua sin(θ + α) √ ∆ub = 2∆u sin(θ + α − 2π ) c(t) (5.45) ∆iD and ∆iF can be expressed in ∆id . ∆id = L′′ ∆id . and ∆iD . 2 ∆iD = − LF LAD − LAD ∆i .46) 2 ∆i = − LD LAD − LAD ∆i .43) 3 ∆uc sin(θ + α − 4π ) 3 These expressions for ∆iD and ∆iF can now be inserted into the expression for ∆Ψd in (5. First. From (5. F d LF LD − L2 AD where the subtransient inductance L′′ in the d–axis has been deﬁned and it d can be rewritten as LD + LF − 2LAD L′′ = Ld − .44) ∆uq 0 steady state Ld and Lq .48) d LF LD /L2 − 1 AD . directly after the voltage is applied on the terminals.72 5. c(t) = 0 for t < 0 and c(t) = 1 for t > 0. there will be couplings between the circuits in Figure 5.43) becomes ∆u0 0 √ ∆ud = 3∆u c(t) (5. With α = π/2 the Park–transformed voltage vector in Equation (5. Synchronous Machine Model For t = 0+. are the currents induced in the circuits due to the step in the voltage.4 and this will result in that the equivalent inductances as seen from the terminals will change. when transients occur. i.g. d 2 LF LD − LAD (5.40) giving ∆Ψd = Ld ∆id + LAD (∆iF + ∆iD ) = Ld − LD L2 + LF L2 − 2L3 AD AD AD LF LD − L2 AD . the subtransient and transient inductances in the daxis will be derived. that is. due to short circuits or other disturbances.e. The function c(t) is a step function at t = 0. These equivalent inductances will be derived in the following. the ﬂux linkages ΨF and ΨD are still zero since they cannot change instantaneously. This gives the following equations ΨF (0+) = 0 = LF ∆iF + LAD (∆id + ∆iD ) ΨD (0+) = 0 = LD ∆iD + LAD (∆id + ∆iF ) (5. However. d (5. ∆id .47) where ∆if . with obvious notation.
which means that after a certain time it can be assumed that the current in the damper winding has decayed to zero.43) becomes ∆u0 0 √ ∆ud = 3∆u 0 (5.40) and iD = 0 ∆Ψd = Ld − L2 AD LF ∆id = L′ id . the eﬀective inductance after the current in the damper winding has decayed is practically equal to the synchronous inductance. since no ﬁeld winding exists in the q–axis for the model we are considering here. d (5. and Subtransient Inductances 73 The time constant of the decay of the current in the damper winding is much smaller that the time constant in the ﬁeld winding.51) (5. Hence it is sometimes said that. Transient. LF (5.52) A similar analysis as above will the give the subtransient inductance in q– axis L2 AQ L′′ = Lq − (5. is deﬁned in Equation (5.e. The same assumption regarding iD can be made if no damper winding is modelled in the daxis. L′ . With α = π the Park–transformed voltage vector in Equation (5. According to the reasoning above we have L′ = Lq . q (5.49) The transient inductance in the d–axis. ∆iD = iD = 0.55) The time constant for the decay of the current in the damper winding is derived in the next section. This assumption gives ∆ΨF = 0 = LF ∆iF + LAD ∆id which gives ∆iF = −(LAD /LF )∆id and with use of (5.4. i. Synchronous.5.53) ∆uq c(t) . for machines with salient poles. For a machine with salient poles and damper winding in the q–axis. but.54) q LQ for a synchronous machine with one damper winding in the q–axis. see next section.51) as d L′ = Ld − d L2 AD . the transient and synchronous inductances in the q–axis are equal. the terminology is somewhat diﬀerent.50) (5. An equivalent analysis to the one above can be performed for the q– axis.
or equivalently i0 = id = iq = 0. If one additional damper winding is modelled into the q–axis. Thus.5 Time constants The time constants that determine how fast transients in the rotor and damper windings decay after a disturbance depend on the resistances of these windings and associated inductances. d d L′′ < L′ < Lq . (This fact was also used above in the derivation of L′′ and L′ . the following relations hold L′′ < L′ < Ld .e. We will here derive the so called open circuit time constants whereby it assumed that the stator circuits are open. which most often are expressed in p. e.58) . e. Since id = 0. see Table 5.e.. for the inductances deﬁned.) Therefore d d we will analyze the decay of the current in the damper winding ﬁrst.1. ia = ib = ic = 0. suppose now that the stator windings are open and and a step change in the ﬁeld voltage at the time t = 0. which also uniquley determine the winding resistances. transient.74 5. Tdo as explained below. In general. based on the machine power and voltage ratings. the ﬂuxes in the ﬁeld and damper windings can be written as ∆ΨD = LD ∆iD + LAD ∆iF . q q The subtransient. The circuit equations describing the currents induced by the step in the ﬁeld voltage for these two windings are for t > 0 ˙ rF ∆iF + ∆ΨF = ∆uF c(t) ˙ rD ∆iD + ∆ΨD = 0 (5. The open circuit time constants are normally the ones given by the manufacturer of the generator and can be obtained by measurements. i. (5.56) 5.u.57) For a synchronous machine rD ≫ rF and the induced current will consequently decay much faster in the damper winding than in the ﬁeld winding.g. with c(t) being a step function at t = 0. Synchronous Machine Model in this case. Xd = ω0 Ld etc. a value of L′ can be derived that diﬀers from Lq . This derivation is analogous to the q one done for the inductances in the d–axis.g. An alternative that also occurs in the literature are the time constants when the stator terminals are shortcircuited. i. We will here derive the time constants associated with the ﬁeld winding and the damper winding in the d–axis. and synchronous reactances are deﬁned in an obvious way. These ′′ time constants are normally denoted by an index o for open. ∆ΨF = LF ∆iF + LAD ∆iD . (5. uF = uF (0) + ∆uF c(t) will be investigated.
e.58) is inserted into the lower equation of (5. when iD = 0. Time constants Taking the time derivative of the lower equation in (5. LF (5. ∆iF . ∆iD . and the transient time constant of ′ the open circuit.58) it is easily seen that this time constant is given by ′ Tdo = LF /rF .5.61) This equation can be rewritten as ˙ ∆iD + rD LAD /LF LAD 2 /L ∆iD = −∆uF L − L2 /L + rF L ∆iF . Tdo . (5. 75 (5. The time constant of the decay of the damper winding current.60) Now the time derivative of the upper equation of (5.63) LD − LAD F F D F AD As pointed out above the current in the damper circuit will decay must faster.60) is inserted giving ˙ rD ∆iD + LD ∆iD + LAD ˙ ∆uF − rF ∆iF − LAD ∆iD =0 . which means that the current in the ﬁeld winding. can be regarded as a constant in Equation (5.58) gives ˙ ˙ ˙ ∆ΨF = LF ∆iF + LAD ∆iD . Tdo . the ﬁeld current is determined solely by the upper equation in (5. and together with the lower equation of (5. .63) as ′′ Tdo = LD − L2 /LF AD .65) Accordingly.66) can be derived.64) When iD has vanished.59) ˙ From the upper equation of (5. is the subtransient time constant of the ′′ open circuit in the d–axis.57) we can replace ∆ΨF in (5. for a synchronous machine with a damper winding in the q– axis.62) (5. ˙ in which the expression for ∆iF in (5.57) yielding ˙ ˙ rD ∆iD + LD ∆iD + LAD ∆iF = 0 . i. (5.63). which after some rearrangements gives ˙ ∆iF = ˙ ∆uF − rF ∆iF − LAD ∆iD LF (5. rD (5.57).5. ′′ Tqo = LQ /rQ (5.59). and can now with the above assumptions be identiﬁed from Equation (5.
) q ′ Tdo (s) ′′ Tdo (s) ′ Tqo (s) ′′ Tqo (s) H (s) Table 5. for several reasons.6 5.5 0.) d x′ (p.12 – 0. these are often the equations used to represent the synchronous machine.u. xd = ω0 Ld etc. simpliﬁed representations of the synchronous machine are often the only possibility. When a synchronous machine is modelled in for example a software package for stability simulations.e.2 – 0.3 0.02 – 0.25 0.15 – 0.6 – 1. i. 5. such a model will be derived under the following assumptions: ˙ ˙ • The stator transients are neglected.05 3 – 5 (n = 3000 rpm) 5 – 8 (n = 1500 rpm) Salient Pole 0.u.u.0 0.6.5 – 5 xd (p.76 5. If a better insight into a problem is wanted or if a problem is analyzed without using computer simulations. we set Ψd = Ψq = 0.4 0.0 0.) q x′′ (p. Typical values of some parameters of synchronous machines. The reasons for this are that they are easily measured by simple tests and that they are introduced in a natural way into the simpliﬁed models we will derive in the next section.2 and 5. In Table 5.09 1. with ω0 = 2πf0 . The quantities introduced in this and the previous sections are important parameters of a synchronous machine and are usually given by the manufacturer of the machine.0 – 2.u.3 1.25 3. base is the power and voltage ratings of the machine. i.u.0 0.4 – 1.05 0.0 – 10.u. Nevertheless.) xq (p.1.1. it is. Here.05 – 0.3 – 1.2 – 0.02 – 0.35 — 0.u. often meaningful to use models that comprise more simpliﬁcations and approximations than those derived earlier.0 – 2.45 1.3. considering the assumptions and approximations made.) d x′′ (p.5 – 2. was presented in sections 5.0 0.01 – 0.1 Simpliﬁed Models of the Synchronous Machine Derivation of the fourthorder model A complete and exact model of the synchronous machine. Reactances instead of inductances are given in Table 5.1.01 – 0.12 – 0. Synchronous Machine Model Round Rotor 1. typical values of the discussed parameters for diﬀerent types and sizes of synchronous machines are given. . The p.0 0.15 – 0.) x′ (p.e..5 0.5 – 9.
68) ud = −rid + ωΨq . 0 = rQ iQ + Ψ (5.5. the stator current. q d (5.75) . i. 77 With these assumptions. L′ = Ld − L2 /LF AD d L′′ = Lq − L2 /LQ q AQ (5.71) can be written as Ψq − e′ /ω = L′′ iq . by deﬁning e′ = ω d Equation (5.73) LAQ ΨQ = L′′ iq . ˙Q . • In the q–axis. from the equations above to get a model that can be used to represent the synchronous machine as a component in a system. (5. except for uF . one damper winding is modelled. Ψq = Lq iq + LAQ iQ .74) is obtained. (5. iQ is eliminated.72) Equivalent derivations can be done for the rotor circuit in the d–axis. From the Equations (5. the synchronous machine is described by the equations Ψd = Ld id + LAD iF . Simpliﬁed Models of the Synchronous Machine • The d–axis contains no damper windings. ΨQ = LQ iQ + LAQ iq . which. and uF . q LQ LAQ ΨQ .67). uq = −riq − ωΨd . which is a control variable that can be changed by the excitation system as described in Chapter 6. LF (5.e. ˙ uF = rF iF + ΨF .67) ΨF = LF iF + LAD id . Ψd − LAD ΨF = L′ id d LF (5. which leads to Ψq − Now. for the ﬁeld winding. LQ (5. That means that the only quantities that should be present in the model are stator voltages.69) For completeness. using the deﬁnition e′ = −ω q LAD ΨF . some relations deﬁned above are repeated here.6.70) The goal is now to eliminate all quantities with indices F and Q.71) (5.
80) can also ˙ be written as ′′ Tqo e′ + e′ − (xq − x′′ )iq = 0 . q d ωLAD uF . Synchronous Machine Model Ψd + e′ /ω = L′ id . rF (5. eq = −ωLAD iF (5.78). ˙ (5.85) (5.78 can be written as 5. i. iQ = e′ − (xq − x′′ )iq q d . ˙d (5. for the exciter winding.76) (5. q d with xd = ωLd etc. accordingly ′ Tdo e′ + e′ + (xd − x′ )id = − ˙q q d ωLAD uF rF (5.78) Substituting this into the second equation of (5. q d eq = e′ + (xd − x′ )id .84) If Equations (5.77) and (5. ωLAQ (5.81) and (5. According to Equations (5.86) . the quantities ed = ωLAQ iQ .82) is obtained.77) are introduced.68). which. can be obtained. From the equations above.84) are Laplace–transformed and the voltage equation (5. (5.e.72) give e′ − (xq − x′′ )iq LQ ′ q rQ d + e =0 . which together with (5.83) (5. q d Now. using eF = can be written as ′ ˙q Tdo e′ + e′ + (xd − x′ )id = −eF . the relationships ed = e′ − (xq − x′′ )iq .79) (5. Equation (5.80) ωLAQ ωLAQ d where it has been assumed that ω = 0 or ω = ω0 . we obtain ∆ud ∆uq with e′ given by d ∆e′ = d = ∆e′ d ∆e′ q + −rd x′′ q −x′ −rq d (xq − x′′ )∆iq q ′′ 1 + sTqo ∆id ∆iq .69) is rewritten with the introduced quantities.81) q d For the rotor circuit in the d–axis.
i. q 5. ′ 1 + sTdo 79 (5. The model shall enable the reader to directly implement a usable representation of an SMIB system. a generic simpliﬁed representation of the excitation system is used. q d Often.2 The HeﬀronPhillips formulation for stability studies The thirdorder model of the synchronous machine derived in the previous section can be formulated as a block diagram. Simpliﬁed Models of the Synchronous Machine and e′ from q ∆e′ = − q ∆eF + (xd − x′ )∆id d . The main cause of the oscillations is a lack of damping in the system due to the eﬀect of excitation systems in synchronous machines. They are known to either decay slowly or continue to grow. This model of the synchronous machine demands.e. providing an additional damping component which leads to an attenuation of the oscillations. Equations (5. lowfrequency oscillations can arise spontaneously. which can be directly attached to the HeﬀronPhillips model derived in this section. By introducing a number of new constants. ﬁeld winding.85)–(5.87) This model.6. will be presented in the chapters 6 and 7.5.87). In the present case. Background: Lowfrequency oscillations in power systems In large interconnected power systems. These are automatic controllers acting on the excitation systems. including the mechanical dynamics. also the damper winding in the q–axis can be neglected. Detailed descriptions and common variants of these systems can be found in chapter 6. and excitation system. The basis for the model presented here. these oscillations have a frequency of well below 1 Hz (about 3 – 10 cycles per minute depending on the power system). a very compact notation is achieved. which leads eventually to system separation. four state variables: ∆e′ and ∆e′ and the “mechanical” quantities ∆ω and ∆δ. . This implementation can be used directly for stability studies. is the ”single machine. are often called the fourth order model.88) with ∆e′ according to Equation (5. which can be written as ∆ud ∆uq = 0 ∆e′ q + −rd xq −x′ −rq d ∆id ∆iq .87). An eﬀective countermeasure to lowfrequency oscillations is the usage of Power Systems Stabilizers (PSS). (5.6. Unlike the electromechanical oscillations described by the swing equation. That leads to a third order model. together with the swing equation. Further details about this system. inﬁnite bus” (SMIB) setup. with the assumptions made here. which was originally proposed by Heﬀron and Phillips.
80 Block diagram 5. where the torque balance ∆Tm − ∆Te is considered as an input and the incremental torque angle ∆δ as an output. . The mechanical system is represented by the system inertia and the damping constant. 2) the eﬀect of the ﬁeld winding (determined q by the ﬁeld winding constant K3 and inﬂuenced by ∆δ over constant K4 ). The electrical part of the system consists of three main parts: 1) the composition of the electrical torque (inﬂuenced by ∆δ over constant K1 and the internal incremental voltage ∆e′ over constant K2 ). The excitation system itself is modeled by q a ﬁrstorder transfer function including the ampliﬁcation factor KA and the time constant TA . Note that all quantities are presented in per unit. Synchronous Machine Model Figure 5. and 3) the eﬀect of the excitation system (inﬂuenced by ∆δ over constant K5 and ∆e′ over constant K6 ).5 presents the transfer function block diagram of the HeﬀronPhillips model.
6.K4 Influence of torque angle on field voltage Voltage setpoint change Tm Mechanical torque change utset 0 KA TA s 1 eF K3 ' Tdo K 3s 1 K2 Te Field winding Influence of internal voltage on electric torque ' eq 1 2Hs System inertia 2 f0 s Synchronous speed Excitation system ut KD Damping constant 5. Simpliﬁed Models of the Synchronous Machine K6 Influence of internal voltage on terminal voltage K1 Influence of torque angle on electric torque K5 81 Influence of torque angle on terminal voltage Figure 5.5. . Block diagram representing synchonous machine in a SMIB system.
91). K3 and K4 have their origin in the ﬁeld voltage equation (5. 2 2 Yd = (C1 X1 − C2 R2 )/Ze . . 1 + ZY = C1 + jC2 . The voltages and current in the SMIB system are related to the admittances and impedances by i = Y ut + Z −1 (ut − u0 ) or Zi = (1 + ZY )ut − u0 . R2 = R − C2 xq . The torque angle δ is deﬁned as the angle between u0 and the internal voltage e′ . d X2 = X + C1 x′ .88) is taken into account with one further simpliﬁcation: the armature resistance is neglected.90) This can be separated into real and imaginary parts. (5. which yields R −X X R where C1 = 1 + RG − XB .92) Now. rd = rq = 0. K5 and K6 come from the equation governing the terminal voltage magnitude. K6 5. C2 = XG + RB .87). Yq = (C1 R1 + C2 X2 )/Ze (5. . While K1 and K2 are derived from the computation of the electric torque. (5. This can be inserted into Equation (5. For the calculations below. Equation (5. R1 = R − C2 x′ . inﬁnite bus” (SMIB) system. X1 = X + C1 xq .93) id iq = C1 −C2 C2 C1 ud uq − u0 sin δ cos δ . (5.91) . which yields id iq = Yd Yq e′ − q u0 2 Ze R2 X1 −X2 R1 sin δ cos δ . vt = vd + jvq . This means that one synchronous generator is connected via a transmission line (represented by series impedance Z = R + jX and shunt admittance Y = G + jB) to an ”inﬁnite” bus with constant voltage u0 . K6 shown in Figure 5. d 2 Ze = R1 R2 + X1 X2 . i.5 describe internal inﬂuence factors within the system and can be found by a comparison of coeﬃcients with the equations governing the synchronous machine dynamics.89) . Below we will calculate these constants from the previously presented thirdorder model. (5. SMIB system setup: The power system under consideration is a socalled ”single machine. . . the armature currents id and iq are assumed to be known.6.82 Derivation of constants K1 . The following deﬁnitions will be used later in the computations: q i = id + jiq . The generator has a terminal voltage ut and injects the armature current i via its terminals.e. Synchronous Machine Model The constants K1 . The current and voltage phasors are deﬁned as shown in Figure 5.
88) with rd = rq = 0.96) Te ≈ Pe = id ud + iq uq . this yields with one further linearization: ∆Te = K1 ∆δ + K2 ∆e′ q . This can be linearized again. (5. .5.6. Phasor diagram of SMIB system. (5. (5. resulting in ∆id ∆iq = Yd Yq ∆e′ + q Fd Fq ∆δ .99) . K1 and K2 from electric torque: The electric torque of a synchronous machine can be approximated by the electric power for small deviations from the synchronous speed. .95) Fq X2 R1 sin δ0 Ze where δ0 is the initial torque angle. (5. Equation (5. (5.98) where the wanted constants K1 and K2 are equal to: K1 K2 = 0 iq0 + Fd Fq Yd Yq e′ q0 (xq − x′ )iq0 d + (xq − x′ )id0 d . Taking into account Equation (5. K6 can be deﬁned in terms of the introduced notation. the constants K1 .97) q d Merging this with Equation (5.96) can be written as Te = iq e′ + (xq − x′ )id iq .94).94) with the constants Fd and Fq introduced for abbreviation. Now. Simpliﬁed Models of the Synchronous Machine 83 qaxis e‘q ut i u0 daxis Figure 5. The overall power is the sum of powers in the d.and qaxis: (5.6. which are equal to u0 −R2 X1 Fd cos δ0 = 2 .
(5.100) Substituting ∆id of Equation (5.107) .94) into Equation (5. (5.103) K5 and K6 from terminal voltage magnitude: The magnitude of the terminal voltage of the generator can be expressed by the d and q components: u2 = u2 + u2 t q d .105) Substituting Equation (5.101) where K3 = 1/(1 + (xd − x′ )Yd ) d K4 = (xd − x′ )Fd d . (5. .84) can be linearized. The corresponding transfer function represents a generic exciter and voltage regulator of the fastresponse type. (5. Synchronous Machine Model K3 and K4 from ﬁeld voltage equation: The ﬁeld winding circuit voltage equation from (5. Excitation system parameters: The only missing parameters for the completion of the HeﬀronPhillips model are the excitation system parameters KA and TA . (5.106) Note that all subscripts 0 denote steadystate quantities in this section.102) (5. which yields ′ (1 + s Tdo )∆e′ = −∆eF − (xd − x′ )∆id q d . (5.104) The deviation from a steady state is thus equal to ∆ut = (ud0 /ut0 )∆ud + (uq0 /ut0 )∆uq . Reasonable values for the parameters are KA ≈ 50 and TA ≈ 0.84 5.05 s.105) yields ∆ut = K5 ∆δ + K6 ∆e′ q where K5 K6 = 0 uq0 /ut0 + Fd Fq Yd Yq −x′ uq0 /ut0 d xq ud0 /ut0 .94) into this yields ′ (1 + s Tdo K3 )∆e′ = −K3 (∆eF + K4 ∆δ) q (5.
cannot be transported over long distances in the system.1 Relation between voltage and reactive power As explained in detail in chapters 2 – 4.1 Normally the power system is operated such that the voltage drops along the lines are small. However. there are several sources and sinks of reactive power. the relation between voltage and the reactive power balance is discussed and various inﬂuences to this balance are presented. On the other hand. 1 85 . Vice versa. The following structure is based on [6] 6. i. while an improper reactive power balance will result in deviations of the voltages in the system from the desired ones. the active power balance of a power system must be maintained in order to keep the system in steady state. the generated reactive power is always equal to the consumed reactive power. overvoltages arise when reactive generation is too high. In this case the transmission system is eﬀectively used. this does not mean that the generated and consumed reactive power is not equal as in the case of active power. If the voltages settle at too low values.e. The frequency deviation is a consequence of an imbalance between power fed into the system by the prime movers and the electric power consumed by the loads and losses. The voltage magnitudes are always adjusted such that this balance is maintained. While the active power is entirely produced by the generators in the system. which is a consequence of Kirchhoﬀ’s laws. If the active power balance is not kept. the reactive power. the reason is that the reactive generation is too small. The primary voltage control equipment in the synchronous machine is also discussed and ﬁnally a short description of the hierarchical voltage control structure is provided. However. First. the frequency in the system will be inﬂuenced. Increased production of reactive power results in higher voltage near the production source. the reactive power balance must be kept in such a way that the voltages are within the acceptable limits.6 Voltage Control in Power Systems Having introduced the principal governing equations for the synchronous machine in the previous chapter. The node voltages of the system will then almost be equal (ﬂat voltage proﬁle). in contrast with the active power. this chapter deals with the basics of voltage control in electric power systems. while an increased consumption of reactive power results in lower voltage. Furthermore. the voltage magnitudes can be controlled to desired values by control of the reactive power. primarily for transmission of active power. Thus. and not for transmission of reactive power.
The reactive power of the synchronous machines is easily controlled by means of the excitation. the reactive power can be regarded as a fairly local quantity. It is most eﬀective to compensate the reactive power as close as possible to the reactive load. and electrical distributions companies to compensate their loads in an eﬀective way. the reactive power is easy to control. The only possibility to radically reduce the total reactance of a transmission line s to connect a series capacitor or a series FACTS device. FACTS devices are power electronics devices equipped with fast control that can often can be used for reactive power control. FACTS = Flexible AC Transmission Systems.86 6. If this percentage is exceeded. For overheadtransmission lines the reactance can be slightly reduced by the use of multiple conductors. e. The high voltage network is in this way primarily used for transmission of active power. Important generators of reactive power are: • Overexcited synchronous machines • Capacitor banks • The capacitance of overhead lines and cables • FACTS devices2 Important consumers of reactive power are: • Inductive static loads • Underexcited synchronous machines • Induction motors • Shunt reactors • The inductance of overhead lines and cables • Transformer inductances • FACTS devices For some of these. Consequently. They can normally be used both as reactive power sources and sinks.g. industries. Switching of shunt capacitors and reactors can also control the reactive power. the consumer has to pay for the reactive power. Voltage Control in Power Systems since normally X >> R in a power system. The reactive losses of power lines and transformers depend on the size of the reactance. These tariﬀs are generally designed so that the reactive power is only allowed to reach a certain percentage of the active power. Further discussion of FACTS devices is given in the Section 6.3.4. while for others it is practically impossible. There are certain high voltage tariﬀs to encourage large consumers. FACTS devices oﬀer also a possibility to control the reactive power. 2 .
In that way.2. since the reactive generation from these is much larger than from overhead lines.2 Voltage Control Mechanisms The following factors inﬂuence primarily the voltages in a power system: • Terminal voltages of synchronous machines • Impedances of lines • Transmitted reactive and active power • Turns ratio of transformers A suitable use of these leads to the desired voltage proﬁle. Voltage Control Mechanisms 87 6.6. the voltage can be kept equal to the set value. Nowadays new installations of synchronous compensators are very rare. These are synchronous machines without turbine or mechanical load. The voltage drop caused by the generator transformer is sometimes compensated totally or partly by this means. Shunt reactors must sometimes be installed to limit the voltages to reasonable levels. and the voltage can consequently be kept constant on the high voltage side of the transformer. Large reactive transmissions cause large voltage drops. In networks which contain a lot of cables this is also necessary. which can produce and consume reactive power by controlling the excitation. A power electronics based device can be economically motivated if fast response or accuracy in the regulation is required. The output from this controls the excitation of the machine via the electric ﬁeld exciter. The generators are often operated at constant voltage. This can be made by the excitation of the synchronous machines. Synchronous compensators can also be installed for voltage control. (C is much larger and X is smaller. and power electronics based solutions are preferred if fast voltage control is needed. there are often no synchronous machines close to the load. so the most costeﬀective way is to use shunt capacitors which are switched according to the load variations. The reactive power transmitted over a line has a great impact on the voltage proﬁle. by using an automatic voltage regulator (AVR). the production of reactive power should be as close as possible to the reactive loads. as described above. Instead. thus these should be avoided.) . However.
PSS. a short description of the functions of the diﬀerent blocks in Figure 6.1 is given: • The exciter supplies the ﬁeld winding with direct current and thus comprises the “power part” of the excitation system. .88 6. load compensation can be implemented if the voltage in a point apart from the generator terminals. • The voltage measurement and load compensation unit measures the terminal voltage of the generator and rectiﬁes and ﬁlters it. which also performs a number of protection and control tasks. such as in a ﬁctional point inside the generator’s transformer. since reactive power is a fairly local quantity. The set values for the voltage controllers are selected so that the desired voltage proﬁle of the system is obtained. the terminal voltage of the synchronous machine is controlled by the excitation system. Voltage Control in Power Systems 6. cf. Below.3 Primary Voltage Control The task of the primary voltage control is to control the reactive output from a device so that the voltage magnitude is kept at or close to the set value of the controller. This oﬀers a very eﬃcient and fast way to control the terminal voltage of the machine.3. Further. The main purpose of the excitation system is to feed the ﬁeld winding of the synchronous machine with direct current so that the main ﬂux in the rotor is generated. Further. Chapter 7. Usual input signals for the PSS are deviations in rotor speed. accelerating power. which is brieﬂy summarized in section 6.4. or voltage frequency. • The power system stabilizer.1. In the following the most important devices for reactive power and voltage control are described. gives a signal that increases the damping to the controller. which in most power system is the most important voltage control. Input signals are pure control signals as well as functions for stabilizing the exciter system.1 Synchronous Machine Excitation System and AVR The reactive power output of synchronous machine can for a given active power level be adjusted within the limits of the capability curve by the excitation system. A schematic picture of a generator with excitation system is depicted in Figure 6. should be kept constant. The selection of set values is the task of the secondary voltage control. Usually the node of the controlled voltage is at the same or very close to the node of the reactive device. 6. • The controller treats and ampliﬁes the input signals to a level and form that is suited for the control of the exciter.
normally containing batteries. Primary Voltage Control Limiter and Protection Voltage Measurements Load Compensation 89 Synchronous Regulator Exciter Machine A PSS A = To the Power System Figure 6. protection. and supervisory functions. often on the same axis as the rotor of the synchronous machine. “black start”.1. Three main types can be distinguished: • DC excitation system. • AC excitation system. over–excitation protection. where the exciter is an AC machine with rectiﬁer. which generator and exciter have. a large number of diﬀerent types of exciter systems is used. Many of these ensure that the synchronous machine does not produce or absorb reactive power outside of the limits it is designed for. Usual functions are current limiters.6. where the exciter is a DC generator. where the exciting current is fed from a controlled rectiﬁer that gets its power either directly from the generator terminals or from the power plant’s auxiliary power system. Schematic picture of a synchronous machine with excitation system with several control. • Static excitation system. . Below. In the latter case. and under–excitation protection. Today. the synchronous machine can be started against an unenergised net. a more comprehensive treatment of some of the functions described above is given. • The limiter and protection can contain a large number of functions that ensure that diﬀerent physical and thermal limits.3. The batteries are usually charged from the net. are not exceeded.
As a rule. the IEEE type DC1 system. but many of these systems are still in operation. Load Compensation Equipment Figure 6. The voltage is measured “somewhat inside” the generator. Voltage Control in Power Systems Uc 1 1 + sT R Σ + U err U c = U T + ( R c + j X c )I T U ref Figure 6. hardly any DC excitation systems are being installed. If a machine is connected with a comparatively large impedance to the system. consisting of a converter for measured values. There are several reasons for the use of compensation in voltage control of synchronous machines.2. Both these signals vanish in steady state. DC Excitation Systems Today.2 shows the block diagram of a compensation circuit. and a comparator. is given in Figure 6. Generally. the models given by the manufacturers and power suppliers must be used. Every manufacturer uses its own design. That is necessary to distribute the reactive power in an appropriate way between the machines. If two or more generators are connected to the same bus. The time . the compensation equipment can be used to create an artiﬁcial impedance between those. One example of a DC excitation system. Block diagram of compensating circuit. it can be said that there is a large number of variants of the diﬀerent excitation systems listed above. typical examples for models will be given. and demands that depend on the application often lead to considerable diﬀerences in the detailed models of the devices in each group. Xc is much larger than Rc . Here. a ﬁlter. This then corresponds to negative values of Rc and Xc . The stabilizing feedback UF is subtracted.90 UT IT 6. The controller is mainly described by the dominating time constant TA and the ampliﬁcation KA . The limits can represent saturation eﬀects or limitations of the power supply. which usually is the case since the generator’s transformer normally has an impedance in the order of magnitude of 10% on basis of the machine. and sometimes a signal from the PSS is added. The input signal for the controller is the voltage error Uerr from the compensation equipment.3. it can be desirable to compensate a part of this impedance by controlling the voltage “somewhat inside” of that impedance. corresponding to positive values of Rc and Xc in Figure 6. In reality.2.
UR . which are described in the sequel. The rectiﬁer of the exciter prevents (for most exciters) the exciter current from being negative. The output signal from the voltage controller. and its characteristic is described by FEX . If saturation is neglected. controls the exciter. The output voltage of the exciter is in this case inﬂuenced by the loading. Some functions have been added. the exciter consists of a smaller synchronous machine that feeds the exciter winding through a rectiﬁer. For shunt excited machines. and its eﬀective ampliﬁcation is 1/KE .6. the eﬀective time constant of the exciter becomes TE /KE . Model of DC exciter system (IEEE Type DC1). The structure of the model is basically the same as for the DC excitation system. These are often small and can then usually be neglected. That constant depends on the synchronous and transient reactances of the exciter. the parameter KE models the setting of the ﬁeld regulator. In Figure 6. To represent these eﬀects. Primary Voltage Control U PSS U RMAX U err 91 + + Σ UF 1 + sT C 1 + sT D U RMIN KA 1 + sT A UR + Σ  1 sT E EF U FE SE ( E F ) + K E sK F 1 + sT F Figure 6. constants TC and TD can be used to model internal time constants in the controller. The exciter consists of a DC machine that can be excited independently or shunt excited. . The term SE represents the saturation of the exciter and is a function of the exciter’s output voltage. that is SE = 0. That distinguishes them from static excitation systems. DC and AC excitation systems are sometimes called rotating exciters. The voltage drop inside the rectiﬁer is described by the constant KC .3.3. which is a function of the load current. AC Excitation Systems For AC excitation systems. The feedback with the constant KD represents the reduction of the ﬂux caused by a rising ﬁeld current IF . since they contain rotating machines. an example of a model of AC exciter systems is shown (IEEE type AC1). EF . the exciter current is used as an input signal in the model.4.
The constant KC represents the relative voltage drop in the rectiﬁer. and a large number of variants exists. The time constants are often so small that a stabilizing feedback is not needed. Sometimes. Voltage Control in Power Systems + Σ  + 1 + sT C 1 + sT D KA 1 + sT A U RMIN UR +  Σ 0 1 sT E UE Π F EX F EX = f ( I H ) EF UF sK F 1 + sT F U FE + Σ + K E + SE K CI F I H = UE IF KD Figure 6. Model of an AC exciter system (IEEE Type AC1). With the latter arrangement. The constant KF can then be set to zero. the exciter winding is fed through a transformer and a controlled rectiﬁer. By far most exciter systems installed today are of that type.5. Since the exciter system is normally supplied directly from the generator bus.4. Static Excitation Systems In static excitation systems. . The primary voltage source can be a voltage transformer that is connected to the generator terminals. the maximum negative ﬁeld current is usually considerably lower than the maximum positive ﬁeld current. but even a combination of voltage and current transformers can be found. Static excitation systems can often deliver negative ﬁeld voltage and even negative ﬁeld current.92 U PSS U RMAX U err 6. it is possible to supplement these voltage sources by using the auxiliary power of the power plant as voltage source. That makes it possible to start the generator in an unenergised net. However. An example of a model of a static exciter system is shown in Figure 6. for example during a ground fault in or near the power plant. the maximum exciter voltage depends on the generator’s output voltage (and possibly its current). This is modelled by the dependency of the limitations of the exciter output on the generator’s output voltage. an exciter current can be obtained even if the voltage at the generator terminals is low.
3. Therefore. It is impossible and also unwise to use the reactive power capabilities of synchronous machines to compensate for this. In systems with long highvoltage lines. Model of a static exciter system.K C I FE 93 U err + + Σ U IMIN 1 + sT C 1 + sT D KA 1 + sT A EF UF U T U RMIN . The varied loading condition of the system during low and peak load situations implies that the reactive power needed to keep the desired voltage magnitudes vary signiﬁcantly.6. when the reactive consumption from loads and line reactances is the highest.3. in most systems. Since these load variations are rather slow and predictable. breakerswitched shunt capacitor banks and shunt reactors are used for a coarse control of reactive power. To keep the voltages at acceptable levels. no fast control is needed and the capacitor banks can be breaker switched. the reactive control capabilities of the synchronous machines are not suﬃcient to keep the voltage magnitudes within prescribed limits at all loading conditions. where Qshunt is the size of the shunt element and Ssc is the shortcircuit power at the node. Primary Voltage Control U PSS U IMAX U T U RMAX .5. while shunt reactors are most often installed in one single unit. shunt reactors might be needed. because of high costs. since many synchronous machines will be driven to their capability limits. The sizes of the reactive shunt elements determine how accurate the control can be. Since typically a line commutated HVDC converter station consumes about 50% as much reactive power as active . A factor limiting the element size is the transient voltage change resulting from switching. the reactive power generation of these lines can be very high during light load conditions. 6.K C I FE sK F 1 + sT F Figure 6. The fundamental frequency voltage change in pu caused by switching a shunt element can be estimated as ∆V = Qshunt /Ssc .2 Reactive Shunt Devices In many systems. and the fast and continuous reactive power control oﬀered by the synchronous machine will not be available. Reactive shunt elements are also used for reactive power and voltage control at HVDC terminal stations. in particular. Generally shunt capacitors are switched on during highload conditions. Capacitor banks. so that the synchronous generators can be used for the fast and continuous control. can often be switched in smaller units.
Certain transformers are equipped with a number of taps on one of the windings. but the same principle can still be applied for voltage control. The voltage at the consumer side can therefore be kept fairly constant even though voltage variations occur on the highvoltage network. the relation between the voltage phasors on the highvoltage side U1 and on the lowvoltage side U2 . (6. 6. Figure 6. the turns ratio of transformer is deﬁned as N1 τ= . When the transformer is loaded. since then lower currents needs to be switched. These ﬁlters are almost purely capacitive at fundamental frequency. by switching out a number of windings on the highvoltage side. since the load current yields a voltage drop over the leakage reactance of the transformer Zk . the taps are placed on the highvoltage winding (the upper side). equation (6.1) N2 Then. If N1 is the number of turns on the highvoltage side and N2 is the number of turns on the lowvoltage side.3. that is. U1 Zk τU2 U2 N1 N2 Transformer with variable turns ratio (tap changer). Time constants in these regulators are typically in the . reactive compensation is needed.3 Transformer Tap Changer Control An important method for controlling the voltage in power systems is by changing the turns ratio of transformers. the voltage on the lower side can be kept constant by decreasing. is U2 = U1 τ . Part of the reactive compensation is usually provided by harmonic ﬁlters needed to limit the harmonic current injection into the ac networks.2) is of course incorrect. Transformer with a variable turns ratio (tap changer). Voltage Control in Power Systems power transmitted. at no load.6.94 6. Voltage control can be obtained by switching between these taps. as illustrated in Figure 6.2) If the voltage decreases on the highvoltage side.6. Transformers with automatic tap changer control are often used for voltage control in distribution networks. (6. Normally. Switching during operation by means of tap changers is very eﬀective and useful for voltage control.
devices using voltage source converters have been introduced. but just manually when the transformer is unloaded.7. which indirectly inﬂuences voltages.3. power electronicsbased equipment. The ﬁrst devices introduced were based on thyristors as active elements. but more recently. The SVC is based on thyristors. the uniﬁed power ﬂow controller (UPFC) is equipped with a shunt part for voltage control. Primary Voltage Control 95 order of tenths of seconds. with voltage levels changing in large steps. Figure 6. and the STATic synchronous COMpensator (STATCOM). the turns ratio cannot be changed during operation.4 FACTS Controllers As previously mentioned. with the latter. can be used for fast voltage and reactive power control. Among these devices. In some transformers. In this case. while the STATCOM is based on a voltage source converter. a faster and more powerful control can be achieved. FACTS devices used for voltage control are connected in shunt. but the main reason for installing these devices is to control active power.6. Thyristor controlled reactor. V is Thyristor controlled reactor.3. 6. voltage variations in the network cannot be controlled. usually referred to as FACTS controllers or devices. . There are also seriesconnected FACTS devices that in principle can be used for voltage control. Two such devices are discussed here: the Static var compensator (SVC). which basically operates as a STATCOM.
96 Static Var Compensator
6. Voltage Control in Power Systems
The SVC may be composed of two diﬀerent shunt elements, that is, a thyristor controlled reactor (TCR) and thyristor switched capacitor (TSC) banks. If fast switching of the capacitor banks is not needed, one can also use breakerswitched capacitors. The TCR is depicted in Figure 6.7 by delaying the ﬁring of the thyristors, a continuous control of the current through the reactor can be obtained, with the reactive power consumption varying between 0 and V 2 /X, where X is the reactance of the reactor. By combining the TCR with a suitable number of capacitor banks, a continuous control of the reactive power can be achieved by a combination of capacitor bank switching and control of the reactor current. Usually, the TCR and TSC are connected to the highvoltage grid through a transformer, as shown in Figure 6.8. The control system of the SVC controls the reactive output so that the voltage magnitude of the controlled node is kept constant. Usually, a certain slope is introduced in the control, as shown in Figure 6.9, which shows the reactive current as a function of the voltage. In the (almost) horizontal part of the curve, around the voltage set point, the SVC control is active. When the SVC has reached its maximum or minimum reactive output, the voltage cannot be controlled, and the device will behave as a pure reactor or pure capacitor; thus, in extreme voltage situations, the SVC behaves as a reactor or capacitor bank. The control of the reactor current is based on thyristors, which limits the bandwidth of the voltage control. If a fast control is needed to compensate for ﬂicker and voltage dips, one has to use technologies based on voltage source converters.
V is
Figure 6.8. Static var compensator.
6.3. Primary Voltage Control
97
Figure 6.9. Static var compensator control.
STATCOM The STATCOM is a device which is based on a voltage source converter. The device is connected in shunt and consists of a capacitor charged with a dc voltage, which provides the input voltage for a voltage source converter, as illustrated in Figure 6.10. The converter feeds a reactive current into the network, and by controlling this reactive current, voltage control is achieved. Since a voltage source converter needs semiconductor elements with current interrupting capabilities, thyristors cannot be used; instead, elements such as GTOs or IGBTs have to be used. In comparison with the SVC, the STATCOM oﬀers two advantages. First, the STATCOM’s output reactive current is not limited at low or highvoltage conditions; rather, the output current is only limited by the converter ratings and is not dependent on the system voltage. This means that the reactive support during extreme voltage situations is much better with respect to the SVC, as shown in Figure 6.11. Second, the control response is much faster, since it is limited by the switching frequency of the voltage source converter (usually around 1 kHz). The STATCOM can hence be used to reduce ﬂicker and other fast voltage variations eﬀectively. The dc capacitor of the voltage source converter constitutes an active power storage; hence, an active current can also be injected into the network. The STATCOM can thus also be used for active power control (example, to damp power oscillations). However, the energy stored in the dc capacitor is fairly small; therefore, to truly control the active power output, a battery must be installed on the dc side.
98
6. Voltage Control in Power Systems
V is
Voltage source converter
Figure 6.10. STATCOM.
V
Capacitive
Inductive
is
Figure 6.11. Voltage control and reactive capability of a STATCOM.
12 illustrates the basic setup. Secondary Voltage Control 99 6. Hence in case of a disturbance. This may result in unacceptable voltage values and an uneven reactive power distribution over the generators. The higher level is responsible for the voltage control of the ’pilot’ node.6. TVR SVR SVR PI PI QR QR QR QR QR AVR AVR AVR AVR AVR Power network Figure 6. balancing the reactive power via AVRs and other devices can be considered a local control action. Structure of the voltage regulation.4 Secondary Voltage Control As previously discussed. a ’pilot’ node is selected and controlled typically by the participating generators so as to maintain the voltage of the pilot node at a speciﬁc value. Figure 6. A similar hierarchy applies also in the time constants of each control level in order to .4. The latter corresponds to a local reactive power regulator (QR) that adjusts the input of the AVR with respect to the generator capability. The voltage error serves as input to a proportionalintegral controller (PI) and the generated signal is then used by each generator at a lower control level. SVR is a hierarchical. This creates the need of a coordinated adjustment of the setpoints of the reactive power suppliers. by delivering the reactive power proportionally to their own capabilities. In each zone. Therefore the AVR is enhanced with two additional control levels. centralized voltage control scheme that supervises the generator AVRs and other reactive power sources in a given network zone so as to enhance voltage stability of the grid. Secondary Voltage Regulation (SVR) has been developed to address the above mentioned situation.12. the devices electrically nearby will try to compensate the reactive power needs.
whereas the time constant of the PI controller is in the order of 50 sec. based on an overall system economic optimization. The described secondary closedloop control scheme is already implemented in some European power grids. the QR time constant is in the order of 5 sec.5 sec. in which the operator adjusts the setpoints manually with an expected performance degradation. The AVR time constant is in the order of 0. Voltage Control in Power Systems avoid any interaction among them. The scheme can be further enhanced with the use of Tertiary Voltage Regulation (TVR) which.100 6. . particularly in France and in Italy (since the early 1980’s). will determine the setpoints of the above mentioned PI controller in a time scale of 15 minutes. Other countries still follow an openloop strategy.
the damping torque is rather small and thus inﬂuences the oscillation frequency only marginally. factors inﬂuencing the stability of electric power systems are discussed. The variation of the synchronizing torque with the rotor angle determines. Diﬀerent sources of positive and negative damping are discussed. either fall out of phase (instability) or oscillate with unchanged amplitude. Furthermore. whereas the damping torque ∆Td is in phase with the deviation in rotor speed ∆ω. strives. When the generator has reached an operating point where the synchronizing power no longer can return the system to the stable equilibrium. together with the machine’s moment of inertia. also called the synchronizing power. The synchronizing torque. and methods to improve the damping are given. If the modes of oscillation in a system are determined by computing the eigenvalues of the linearised system’s Jacobian matrix. the frequency of rotor oscillations. This is not realistic. the generator will fall out of phase.7 Stability of Power Systems In this chapter. 7. the eﬀect of the loads on system stability is regarded in the second part of this chapter. The partitioning into synchronizing and damping torque is shown in Figure 7.1. to bring the rotor back to the stable equilibrium in which the mechanical power is equal to the electrical power. since real systems contain damping.1.1 7. Normally. Normally. a short introduction to damping in a power system is given. Therefore. First. after a disturbance. two diﬀerent kinds of electrical torques appear at a generator rotor that is oscillating: a synchronizing torque ∆Ts and a damping torque ∆Td .1 Damping in Power Systems General What damping in the context of electro–mechanical oscillations in a power system means is quite self–evident. changes in the synchronizing and damping torques will become apparent as follows: An increase of the synchronizing torque moves the eigenvalue parallel to the 101 . if it is positive. The synchronizing torque ∆Ts is in phase with the deviation in rotor angle ∆θ. The damping torque depends on the time derivative of the rotor angle in such a way that the oscillation is damped. It mainly inﬂuences the amplitude. the system will. In a synchronous machine. the main contributors to damping are the damper windings and the ﬁeld winding. Damping is neglected in the classical model.
some loads contribute with positive damping.1. and hence. 7. all eigenvalues will be situated on the imaginary axis. Necessary for stability is that no eigenvalues are situated in the right half plane. This can. (Earlier. If the damping torque is increased instead. necessitate tripping of lines and it must be avoided.1. That damping is determined by phases and amplitudes of the oscillating torques caused by induced currents in exciter winding and damping windings. Stability of Power Systems ∆ω ∆T d ∆T e ∆T s ∆θ Figure 7.) A common reason for low damping is the use of voltage controllers with high gain. In the classical model. imaginary axis towards larger values.1. This corresponds to positive ∆Ts as well as positive ∆Td in Figure 7. the eigenvalue will move parallel to the real axis to the left. Partitioning of electrical torque in synchronizing and damping components. Further. Before the reason behind this phenomenon was known. that type of instability was called dynamic instability. These contributions originate from the frequency dependency of the loads. It can be shown that an eigenvalue with positive real part can occur when large amounts of power are transmitted and voltage controllers with high gains are used. in the worst case. for very slow oscillations. Such a conﬁguration can also be analyzed comparably easily. This corresponds to an increase in the spring constant in a mechanical analogy.102 7. but also their voltage dependency contributes. That was experienced in generators feeding a strong net through a line. That type of instability is called small signal instability or instability caused by low damping. the problem was solved by operating the generator with manual voltage control. Generally. or by . the internal damping of a generator comes from the windings in the rotor circuit.2 Causes of Damping As mentioned earlier. The currents in the damping windings decay. Low or negative damping in a power system can lead to spontaneous appearance of large power oscillations. their contribution is small. the inner damping of the generators decreases with decreasing frequency of the oscillations.
The constant Kstab determines the size of that contribution. If the ampliﬁcation in the voltage controller is high. making the voltage controller slower or decreasing its gain. it can be said that the rotor angle inﬂuences the generator voltage. such as HVDC (High Voltage Direct Current) or SVC (Static Voltage Condensers). in this case ∆ω. The operating principle for these is very simple. Now it turns out that. That constant should of course not be chosen larger than necessary to obtain the needed damping. Other possibilities for increasing the damping in a system are diﬀerent types of controllable equipment that may be installed in the system. when the load on the machine is high. ﬁrst passes a high–pass ﬁlter to ensure that permanent frequency deviations do not contribute. the same physical mechanism in the system of generator and voltage controller that above resulted in negative damping is used to obtain positive damping.1. in the generators. but they are usually too expensive to install them only to increase the damping. The next ﬁlter shifts the phase appropriately for the critical oscillation frequency so that a positive contribution to damping is obtained. These components can often give large contributions to damping. the phase angle can be such that a contribution with negative damping is obtained. Such a power system stabilizer usually utilizes the rotor deviation from the synchronous frequency ∆ω as input signal. like Pe or Te . A diagram illustrating the principle mode of operation of a PSS is given in Figure 7. Damping in Power Systems ∆ω sT W 1 + sT W 1 + sT 1 1 + sT 2 103 ∆U ref K STAB Figure 7. which inﬂuences the electrical torque. a signal is added to the reference voltage of the generator’s voltage controller. Thus. other signals that contain the same information can be used. The phase of this signal should of course be such that it results in a positive contribution to the damping. The input signal.7. Sometimes. that negative contribution can be signiﬁcant. since this could lead to undesired side eﬀects.2.1. The simplest and usually cheapest way is the installation of power system stabilizers. and the existing equipment is not always located optimally for damping purposes. To explain that mechanism in detail is beyond the scope of this compendium. PSS. Block Diagram of a simple PSS. 7. To increase the damping in the system.3 Methods to Increase Damping Several methods for increasing the damping in a power system are available. which through the voltage controller inﬂuences the transient emf. but in summary.2. .
sometimes diﬀerent load models have to be used at diﬀerent buses. Sometimes. that is.2. deriving models for single load objects is formally not very diﬃcult. since it varies during the day as well as during the year. lumped loads as they are perceived from a bus in the high voltage grid. loads are usually modelled with static models. The same is true for their voltage dependency since it inﬂuences the voltage control. The voltage characteristics of the loads have a direct inﬂuence on the accelerating power for generators nearby and are thus very important for the behaviour during the ﬁrst oscillation after a fault. The most common load models are presented. Voltage stability is highly dependent on the balance of reactive power in the system.1. an equal amount of power is consumed in the loads in the system as is generated in the generators. The frequency dependency of the loads also inﬂuences the system damping. It is obvious that the voltage dependency of the loads is of high importance for the system’s voltage stability. This chapter discusses brieﬂy how load characteristics inﬂuence the system stability and which problems arise in the derivation of appropriate load models. (Of course. This is a measure for the voltage stability of the system. This compendium concentrates on what is usually called angular stability. the load characteristics are on principle just as important for the system properties as the generators.2 Load Models For studies of angular stability. It is for several reasons diﬃcult to derive good load models. the ability of the generators to stay synchronized after disturbances. Stability of Power Systems 7. domestic loads. it is diﬃcult to estimate the composition of the loads. large induction motors have to be represented individually .104 7. 7.2. neglecting losses. depending on the composition of the loads. Thus. but also the active power has some inﬂuence here.2 that the frequency dependency of the loads inﬂuences directly how large the frequency deviation after diﬀerent system disturbances will become. 7.2 Load Modelling Since. Further. and rural loads. Another important property of a power system is the ability to keep the voltages in the system within acceptable limits during disturbances. That is. or synchronous stability.1 The Importance of the Loads for System Stability The characteristics of the loads inﬂuence the system stability and dynamics in many diﬀerent ways. this composition varies from bus to bus. not reﬂected in the level of detail and the accuracy usually used in load models for analyzing system stability. It has been shown in section 2. for example industrial loads.) First. Loads here are. however. however.
e. Q = 0. it is called constant current load. The ZIPModel is a special case of the model in Equation (7. like rotor oscillations. Some examples for the voltage exponents of diﬀerent loads are: Electric heating: α = 2. a somewhat larger voltage exponent is used for the reactive load. the above models have to be modiﬁed. the most commonly used model types are static models. It is very common to model the (active) load for stability studies as consisting of these three parts. (7. Usually.1). it is a constant impedance load. However. A load model often used is the so called ZIP–Model. where the investigated time frame is at most around 10 s after the disturbance. To be able to include active and reactive losses in the underlying distribution grid. the load is called constant power load. constant impedance (Z) load . (7. U0 is the nominal voltage at nominal load. 9. The most common model for voltage dependency is P = P0 U U0 U U0 α . They are called static since they describe the load using only algebraic equations. the load can also be modelled as an arbitrary polynomial in (U/U0 ).2). Fluorescent tubes: α ≈ 0. β ≈ 2.1) β Q = Q0 . 6. which contains three terms with: • α1 = 2. Q = 0. and if α = 2. Dynamic load models for lumped loads have begun to be used during the last few years. especially for studying voltage stability. that are modelled.2). In Equations (7. i. U αi P = P0 ki . Light bulbs: α ≈ 1. Load Modelling 105 by special models to obtain the correct dynamic behaviour. if α = 1.2) U0 i In Equations (7. Static Load Models For traditional stability studies. If α = 0. P0 and Q0 . the voltage exponents α and β are often diﬀerent. but those are expected to be used in the future more widely and even for other types of studies.1) and (7. The modelled load dynamics are in these cases so fast that they can be considered instantaneous compared with other phenomena.2.7.
The value of L′ can be obtained from the equivalent circuit of the induction motor shown in Figure 7. This is the basis for the function of the induction motor. . the load in certain nodes is dominated by electric motors. It can then be justiﬁed to model those explicitly. a current that generates a ﬂux is induced in the exciter coil.3.e. A large part of the motor load consists of induction motors that can be modelled as follows: An induction motor is basically a synchronous machine with short–circuited exciter coil. it can be necessary to use a more accurate representation.3. Synchronous machines are then modelled according to the models derived in Chapter 5. with the mechanical part Pm depending on the characteristic of the mechanical load. Between the rotating synchronous ﬂux generated in the phase windings and the ﬂux from the exciter winding. according to Figure 7. Stability of Power Systems X′ = ωL′ + e′ rs i + e Figure 7. Equivalent circuit for induction motor. energy is exchanged. a motor load behaves approximately like a constant power load. Lr rr s Lm Ls rs Figure 7.4. The induction motor can. constant current (I) load • α3 = 0. i. Sometimes. be described by a voltage source behind an impedance. For larger voltage changes.106 7. constant power (P ) load Motor Loads Around half of all electric power used by the industry is used for operation of motors. For small changes in voltage. • α2 = 1.4. If the exciter coil rotates with an angular speed diﬀerent from the rotating ﬂuxes generated by the three phase coils. Representation of induction motor.
Since this control exists at diﬀerent voltage levels (cascade coupled controllers). If phenomena taking place in a longer time frame should be studied. the control has to be slower the lower the voltage level is. ω is the machine’s angular speed.6) (7. as mentioned. Lm is the magnetizing inductance.4 rs and Ls are the stator resistance and inductance. Load Modelling In Figure 7. slow dynamics in the system have to be accounted for. Lm + Lr 107 (7. Generally. dt τ0 i= e − e′ . typically in intervals of some tens of seconds. ′ τ0 = (Lr + Lm )/rr are no load operation constants. Equivalent Dynamic Loads The load models presented above are.5) (7. The slip s is deﬁned by ω0 − ω s= ω0 and thus L′ is given by L′ = Ls + The dynamics are described by de′ 1 + ′ (e′ + jω0 (Ls + Lm − L′ )i) + jω0 se′ = 0 . but common to most systems are that tap changers are stepped. The control of tap changers can be done in several diﬀerent ways. Tl is the load torque.4) (7. rs + jω0 L′ Lm Lr . These dynamics originate mainly from two diﬀerent sources: The tap changers installed at lower voltage levels that try to restore the voltage to the desired value and the controllers installed at the loads. . valid for studying phenomena that do not last longer than about ten seconds after a disturbance.3) (7. ω0 is the system’s angular speed.7) (7. rr and Lr are rotor resistance and inductance. dt 2Hm Te = ℜ(e′ · i∗ ) . Here.2.7. until the voltage is restored .8) dω 1 = (Te − Tl ) . undesirable overshoots in the control can occur if the control loops are not coordinated.
Hence. For the reactive load. U = U (t) in the equations above. Measurements of load characteristics have during the last few years become very important. Pl (t) is the value of the active load at the time t. it can be said that the dynamics determined by tap–changer control and load dynamics are highly complicated. dt Pl (t) = Pr + Pt (U ) . It is clearly visible how the load drops momentarily. to recover later to a considerably higher level. (7. A typical example of a load behaviour after a voltage drop is shown in Figure 7. Measurements are needed to get reliable results. The changes in this type of load are determined by thermostats.108 7.5. Summarizing. consists in many areas of heating loads.2. as described by the load models from Section 7. That time is determined by the thermal time constants for what is heated. Of course. it takes some time until. for example. A rather general description is given by Tp dPr + Pr = Ps (U ) − Pt (U ) . Pt (U ) is a static model for the transient load behaviour. and much work is being done in many utilities to investigate load characteristics under diﬀerent loading conditions. Ps (U ) is a static model for the long term load behaviour. such as houses.10) . Stability of Power Systems In Sweden. where Pr (t) is a state variable.2.9) (7. and by the design of the thermostats. at least in winter. a large part of the load. a voltage drop becomes apparent. similar behaviour and equations are valid.
Load Modelling 109 1.9) and (7.5.95 ∆P t ∆P . The Transient Behaviour of the Load in Equations (7.10) after a Step in the Voltage.85 0.8 0.7.9 PRF = 0.∆P s t ∆P t 0.2.05 Load Recovery Factor Definition 1 ∆P s Active Power [pu] 0. .75 5 0 5 10 Time [minutes] 15 20 Figure 7.
Stability of Power Systems .110 7.
2009. GomezExposito. Power Appar. Vittal: Power System Analysis. IEEE Trans. Nov. McGrawHill Inc. [4] W. Academic Press. 692 – 697. Am. Phillips: Eﬀect of modern aplidyne voltage regulator on underexcited operation of large turbine generators. 111 ..G.References [1] P. Inst. Trans. Prentice Hall. [6] A. 1973. As cited in: YaoNan Yu. 1983. ETH Zurich. 1952. [5] A. New York. [3] G. 71. Eng. Kundur: Power System Stability and Control. 1994 (ISBN 007035958X). 2000. A. 19041915. Bergen and V. 2009. J. CRC Press. Canizares: Electric Energy Systems: Analysis and Operation. Heﬀron and P. Second Edition. Electr. [2] Dynamic Models for Steam and Hydro Turbines in Power System Studies./Dec. Electric Power System Dynamics. Andersson: Modelling and Analysis of Electric Power Systems. Syst. Part 3. Conejo and C.A.R..
112 REFERENCES .
(A. d2 θ J 2 = ∆T .u. Tbas.Appendix A Connection between per unit and SI Units for the Swing Equation If nothing else is given after a quantity.u.6) . it is assumed that the nominal electrical and mechanical frequencies are equal.3) and (A. ω0 (A. When using electrical degrees.1) can be written as ∆T = ∆T (p. ω = angular velocity (rad/s). Si = rated power of synchronous machine.i = leading to Si . θ = angle (rad).).4) Si . p. that quantity is in SI–units. π The Hfactor. ∆T = eﬀective torque on the rotor turbine (Nm).u) . (A. The per unit base for torques at synchronous machine i. ω0 Using (A.1) is usually written as M d2 θ · = ∆T . (rad/s). for synchronous machine i is deﬁned by 1 Jω 2 Hi = 2 0 . ω0 dt2 (A.1) dt with J = moment of inertia for rotor turbine (kgm2 ). (A.5).5) (A.2) with M = moment of inertia = Jω0 180 (Js/el◦ ). For simplicity. or constant of inertia. If a quantity is expressed in per unit. 2 dt ω0 ω0 113 (A. In SI–units.) 2Hi Si d2 θ Si · 2 = ∆T (p. is given by Tbas.u. is given in brackets after the quantity (p.i .3) Si with Hi = constant of inertia for synchronous machine i (s). Equation (A.
) = ˙ or ω= ˙ ∆P (p.12) For rotor oscillations.13) can be approximated by ω(p.u.u) .u.u.14) and (A. 2Hi (A. 2Hi ω(p.9). 2Hi (A. (A. .)) = dt 2Hi P = T ωm .) (A.14) ω0 ∆P (p. Of these.) . Equations (A.9) and (A. according to the assumptions. With the equations above.10) (A.114 A. and (A. (ω(p.u) . Connection between per unit and SI Units for the Swing Equation or d2 θ ω0 = ∆T (p.13) (A. this gives P (p.u) ω(p. dt 2Hi or d ∆T (p.u) = T (p.15) are good approximations as long as ω ≈ ω0 .8) (A.9) (A.7) can also be written as d ˙ ω0 (θ) = ∆T (p.10) are exact if ω is the actual angular frequency. 2 dt 2Hi (A. That is valid for “normal” oscillations in power systems.u) .10) now becomes ω(p.u.10).14). (A.15) The most common equations in literature are (A.u) .) .11) Generally. Equation (A. with the actual mechanical angular speed of the rotor ωm that.u) . is equal to the electrical angular speed ω. (A.) ≈ 1 and (A. (A.15).u. ω(p.) 1 · .7) Equation (A.u.) = ˙ ∆P (p.u. dt 2Hi which is the same as d ω0 (ω) = ∆T (p.u.
i. However. but since the relative motion determines the voltage in the phase windings. That ﬂux induces a voltage in the phase winding that is given by U (t) = dΦ ˆ = −Φ0 ω0 sin ω0 t = −U sin ω0 t . the rotor will start to oscillate relative to an undisturbed reference rotor that continues to rotate with the angular speed ω0 . For simplicity. consider a single phase synchronous machine according to Figure B.e.2) Now. A three phase machine has two more phase windings shifted ±120◦ relative to the phase winding in the ﬁgure. so the assumption above does have a practical background. From now on. dt (B. If the balance between power into the generator and power from the generator. between mechanical torque and electrical power. it is even possible to think of stationary ﬁeld windings and rotating phase windings. In all modern larger synchronous machines. almost all relationships and conclusions are independent of this assumption. (B. it is assumed that the ﬁeld winding is in the rotor. the resulting induced voltages in the phase windings will be shaped ideally like a sinusoid. The rotor position can generally be described by θm (t) = ω0 t + θ(t) .Appendix B Inﬂuence of Rotor Oscillations on the Curve Shape If the relative movement between the ﬁeld winding of a synchronous machine and its phase windings is a purely rotating motion with constant angular speed.1.1) with the angular speed of the rotor ω0 according to the system’s electrical frequency. we shall study how the ﬂux linkage through the phase windings will be inﬂuenced when rotor oscillations appear in the system. (B. is disturbed. the ﬁeld winding is on the rotor. while the phase windings are on the stator. The phase winding and the exciter winding are arranged so that the ﬂux linkage through the phase winding is sinusoidally shaped as a function of the angle θm in Figure B. It has earlier been mentioned that stable solutions of the swing equation for a synchronous machine connected to a strong grid consist of oscillations that are nearly sinusoidal with frequencies on the order of magnitude of some tenths of a Hertz to some 115 .3) where θ(t) is a solution of the swing equation.1: Φ(t) = Φ0 cos θm = Φ0 cos ω0 t .
(B. The momentary angular frequency. as given by 1 π Jn (µ) = cos(µ sin x − nx)dx .5) which implies that the oscillatory movement contains a phase–angle modulation of the ﬂux linkage. Schematic picture of single phase synchronous machine. Inﬂuence of Rotor Oscillations on the Curve Shape θm N S Figure B.116 B.6) dt and varies between ω0 + µωr and ω0 − µωr . To investigate how the linked ﬂux. (B.9) .4) with the angular speed ωr corresponding to the oscillation frequency and the amplitude of the oscillatory movement µ.1. and thus the voltage and the current. look during an oscillatory movement. The ﬂux linkage can now be written as Φ(t) = Φ0 cos θm (t) = Φ0 cos(ω0 t + µ sin(ωr t + θr )) . (B.8) π −π An important property of Jn (µ) that will be used later is J−n (µ) = (−1)n Jn (µ) . text books on modulation theory) that Equation (B. is assumed.5) can be written as n=∞ Φ(t) = Φ0 n=−∞ Jn (µ) cos((ω0 + nωr )t + nθr ) . Hertz. is deﬁned for Ψ(t) as d Ω(t) = (ω0 t + µ sin(ωr t + θr )) (B.7) Jn (µ) is a Bessel function of the ﬁrst kind with the argument µ and degree n. It can be shown (cf. a rotor motion according to θm (t) = ω0 t + µ sin(ωr t + θr ) . Ω(t). (B. (B.
Generally. say less than µ ≈ 0.1 0.8 0.1 0 0 0.7) follows that. In Figure B. Figure B.5 0.6 0. If the amplitude is very small.2.117 1 0. which in this context is quite a large amplitude. This justiﬁes the usual representation of the grid with the traditional phasor model with a constant frequency corresponding to f0 .2 0.e. lower than 3 Hz. Φ(t) can thus be approximated quite accurately by n=2 Φ(t) ≈ Φ0 Jn (µ) cos((ω0 + nωr )t + nθr ) . corresponding to an amplitude of approximately 10◦ .8 0.9 1 J2(µ) J3(µ) J1(µ) J0(µ) Figure B. n=−2 (B. From Equation (B.7 0.7 0.10) Typical oscillation frequencies are.4 0.5 µ 0.6 0. in spite of that the momentary angular frequency for Φ(t) is between ω0 +µωr and ω0 −µωr . values for the ﬁrst four Bessel functions are shown for the argument µ between 0 and 1.2.2. side bands with n = 2 and higher can . in most cases considerably.3 0. the coeﬃcients Jn (µ) decay rapidly when the order n becomes larger than the argument µ. 50(60) ± 6 Hz.4 0. A relevant question is how large the amplitudes of these side bands are.2 shows that side bands with n = 3 and larger can be neglected even for amplitudes as large as µ = 1. i. so that 1 corresponds to approximately 57◦ .9 0.3 0. Φ(t) will have inﬁnitely many side bands with the frequencies ω0 ± nωr beside the fundamental frequency ω0 . so that practically the whole energy spectrum for Φ(t) lies in the frequency area f0 ± 6 Hz.2 0. Bessel functions of the ﬁrst kind of order 0 to 3. It should be observed that µ is measured in radians.
including generator rotor and. To increase the understanding for the partitioning in Equation (B. exciter. be neglected. the subsynchronous frequency.12) Equation (B. Inﬂuence of Rotor Oscillations on the Curve Shape ωr µ 2 µ 2 ωr ωr 1 Figure B. according to the text. for ω0 ± ωr . 2 are valid. J0 (µ) ≈ 1 − µ J1 (µ) ≈ . see Figure B.2. i.3. (For an axis with n distinct “masses”. Vector with superimposed oscillation.12). the frequency can deviate signiﬁcantly from the nominal frequency.9) has here been used. maybe.) For the side band in Equation (B. Further.118 B.12). the supersynchronous frequency. it can be shown that.11) . The frequencies of the natural oscillations on that axis are typically 5 Hz and higher. The most common cause of SSO are torsional oscillations on the axis connecting the turbine(s) and the generator rotor.e. If there is a resonance phenomenon at a subsynchronous frequency it is called subsynchronous resonance. the partitioning according to Equation (B. That approximation is used even in a context of power system analysis other than rotor oscillations. for example. series compensation. Since the electrical damping for the lower side band. 4 2 2 (B. can be negative. a more intuitive derivation than the stringent mathematical one using Bessel Φ(t) ≈ Φ0 ((1− µ2 . there are n − 1 diﬀerent eigenfrequencies. SSR. is almost always positive. namely when studying so–called subsynchronous oscillations. and Φ(t) can be written as µ2 µ µ ) cos ω0 t + cos((ω0 + ωr )t + θr )− cos((ω0 − ωr )t − θr )) . The damping in the upper side band. due to. if µ is small. 4 (B. so that it is normally not possible to look only at the component with nominal frequency. SSO.12) has to be kept.
Two vectors with amplitude µ/2 rotate with the angular frequency ±ωr according to Figure B. . Consider a vector with amplitude 1 that performs small oscillations with an angular frequency ωr and amplitude µ. One vector does not move and lies along the horizontal axis. That vector can now be partitioned into the three vectors with unﬁlled arrowheads. Equation (B. Since the vectors rotate with the angular frequency ω0 with respect to a stationary system.3. This can be illustrated geometrically according to Figure B.3. The vector with ﬁlled arrowhead oscillates symmetrically around the horizontal axis with the frequency ωr and the amplitude µ.119 functions can be given.12) is obtained directly from the projection of the vectors on to the horizontal axis. with the modiﬁcation that the factor for the fundamental frequency is one. It is easily observed that the sum of the vectors with hollow arrowhead is at all times equal to the vector with ﬁlled arrowhead.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.