Power System Stability

Prof. Alberto Berizzi Dipartimento di Energia

Ph. 02 2399 3728
Email: [email protected]
 What is stability? 2

• All states of a power system are dynamic states, because some

derivatives of the state variables are always nonzero: x  f ( x, u )
• Let x1(t) and x2(t) be two trajectories of a dynamic system corresponding
to some initial conditions. The system is stable in a Lyapunov sense if,
for any t0, it is possible to identify a number  such that, if

x 2 (t0 )  x1 (t0 )    x 2 (t )  x1 (t )  
holds for any t0  t  
• If, when t tends to infinity, trajectories 1 and 2 tend to be the same, the
system is said asymptotically stable
• The above definition holds for any trajectory, in particular for a trivial
trajectory such as an equilibrium point x̂

Machowski, Bialek, Bumby:

Power system Dynamics

© Alberto Berizzi - Dipartimento di Energia

Power system stability classification 3

• Stability is the ability of an EPS, for a given initial operating condition,

to regain a state of operating equilibrium after being subjected to a
physical disturbance, with most system variables bounded so that
practically the entire systems remains intact

• In AC systems, all rotating machines must be in synchronism

• Any perturbation able to make one or more machines to lose
synchronism can endanger the entire power system
• This is the problem of rotor angle stability
• Other stability problems can be:
 Frequency stability: general imbalance of power, driving the
system to unacceptable frequency values
 Voltage stability: lack of reactive resources that can result in a
voltage collapse

• All mechanisms might cause blackouts more or less spread

© Alberto Berizzi - Dipartimento di Energia
Disturbances 4

• So far, we studied systems at steady-state  algebraic equations

• Disturbance: when a power system undergoes a sudden change or a

sequence of changes (load increase/decrease, generation
increase/decrease, operation of breakers, etc.)

• We can study the consequent steady-state, by a PF (algebraic

equations), but we need to know if that new steady-state will be
reached or not: the problem now involves the dynamics of the system
and must be studied according to its differential equations

© Alberto Berizzi - Dipartimento di Energia

Classification of disturbances 5

• Large disturbance: when the nonlinear equations (algebraic and

differential) describing the power system behaviour cannot be
linearized
 Short-circuits, large load changes, loss of generation units
 The corresponding stability can only be studied by using the
complete dynamic simulation, taking into account every
nonlinearity, like protection systems, opening of breakers, etc.
 Stability depends both on the perturbation and on the system

• Small disturbance: when the same nonlinear equations can be

linearized and studied without loss of accuracy
 Change of the gain of a regulator, small load change, etc.
 Linearization can take place, and the corresponding stability is
dependent only on the intrinsic system features, not on the
perturbation

© Alberto Berizzi - Dipartimento di Energia

Different time frameworks (IEEE) 6

© Alberto Berizzi - Dipartimento di Energia

IEEE Classification of stability (2021) 7

© Alberto Berizzi - Dipartimento di Energia

 Frequency stability 8

1000 MW perturbation in Spain (frequency recorded in Spain, Sicily, Milano)

© Alberto Berizzi - Dipartimento di Energia

 Frequency stability (The Italian job, 2003) 9

Monfalcone=-327MW
50,5 Entracque+S.Massenza+S. 0,1
Fiorano+Presenzano=2159
50,0
0,0
49,5
Montalto+Trino+Livorno=-677MW
ISAB=-180 MW
Bargi+Roncovalgrande+Anapo=1157
49,0 -0,1

df/dt [Hz/s]
La Spezia=-622 MW
48,5
f [Hz]

-0,2
Edolo=538
48,0
TVDN=-502 MW
47,5 -0,3
P.Tolle=-400 MW Trino=-105 MW
47,0
P.Tolle=-496 MW Trino=-105 MW -0,4
46,5

46,0 -0,5
03 30

03 40

03 50

03 00

03 10

03 20

03 30

03 40

03 50

03 00

03 10

03 20

03 30

03 40

03 50
00
5:

5:

5:

6:

6:

6:

6:

6:

6:

7:

7:

7:

7:

7:

7:

8:
:2

:2

:2

:2

:2

:2

:2

:2

:2

:2

:2

:2

:2

:2

:2

:2
03

Time

© Alberto Berizzi - Dipartimento di Energia

Rotor angle stability 10

• Rotor angle stability refers to the ability of synchronous machines of

an interconnected power system to remain in synchronism after being
subjected to a disturbance.
• It depends on the ability to maintain/restore equilibrium between
electromagnetic torque and mechanical torque of each individual
synchronous machine in the system
• Instability that may result occurs in the form of increasing angular
swings of some generators leading to their loss of synchronism with
other generators

© Alberto Berizzi - Dipartimento di Energia

Electromechanical oscillations 11
(Small perturbation angle stability)

© Alberto Berizzi - Dipartimento di Energia

Dynamic security assessment (DSA) 12

• DSA includes:
 Rotor (power) angle stability, voltage stability, frequency stability
 Frequency excursion beyond specified thresholds during
transients
 Voltage excursion beyond specified thresholds during transients
 Damping of power swings inside subsystems and between
subsystems
• Computer programs have to process a huge amount of data
• Operators must receive a limited number of significant results,
preferably graphically
• Preventive and corrective actions must be computed

© Alberto Berizzi - Dipartimento di Energia

Large-disturbance angle stability 13

• Large-disturbance angle stability is typically related to the

behaviour of one or more synchronous machine following a load loss.
This is also called first-swing transient stability or, improperly,
transient stability and involves the study of the first 1-2 s after the
perturbation
• The dynamic model to be adopted depends on the goals of the
analysis and on the time constants involved
 Sometimes you need to take into account excitation systems,
governors, protections, …
 The goal is to determine whether, after a suitable time, the
machine comes back to synchronism or not

© Alberto Berizzi - Dipartimento di Energia

Mechanical analogy 14

• Let’s assume a string is

suddenly cut. Then two
possible end effects are:
 The system will settle down
to a new equilibrium after
some damped oscillations
 Due to transient forces, one
more string will break,
causing a still weaker
network, etc.
• If the system survives is
said to be transient stable
for the considered fault

© Alberto Berizzi - Dipartimento di Energia

Fault close to a generator
15

Fault clearing in 0.25 s

© Alberto Berizzi - Dipartimento di Energia
 Fault close to a generator 16

Fault clearing in 0.34 s

The final part of the transient is determined by
electromechanical oscillations
© Alberto Berizzi - Dipartimento di Energia
Fault close to a generator 17

Fault clearing in 0.35 s

© Alberto Berizzi - Dipartimento di Energia

 Large-disturbance angle stability 18
assumptions
• Only synchronous frequency currents and voltages are considered
in stator windings and power system: DC offset currents and harmonics
are neglected
• Generators are dealt with individually, different from frequency control.
• The frequencies of individual rotors are very close to the initial frequency
f0. Therefore, network equations based on phasors can still be
considered valid: for the grid, algebraic equations are used
• Symmetrical components are used in case of unsymmetrical faults
(however, three-phase faults only are considered, usually), where only
positive sequence currents or voltages will result in synchronizing forces
within the machines
• Generator voltage magnitudes are assumed unaffected by speed
changes
• Generator dynamics are studied, in a simplified approach, by the swing
equation
• Accordingly, the phasor algebra is adopted for network equations,
and dynamics are only considered for rotating machines: DAE
© Alberto Berizzi - Dipartimento di Energia
The mechanical swing equation 19

• The swing equation is the basic equation that drives the motion of the
rotor of a generator
• The accelerating torque (difference between the shaft torque Tm and
the retarding or opposing torque Te) is the product of the momentum
of inertia (including the prime mover mass) times its angular
acceleration (m=mechanical)

d 2m
J 2
 Ta  Tm  Te
dt
• The equation is based on a fixed reference for angles. It is more
convenient to adopt a reference rotating at synchronous speed sm
dm d m d m d  m
2 2
m   smt   m    m   sm   2

dt dt dt dt 2
d 2 m all absolute values
J 2
 Ta  Tm  Te
dt
© Alberto Berizzi - Dipartimento di Energia
The mechanical swing equation cont’d 20

• We can express the swing equation in terms of power [W]:

d 2 m
J m 2
 Pa  Pm  Pe
dt
• Note that in this equation all dissipative terms are neglected
• If the angular momentum is defined, as

• Note that M [J s rad] is constant, because it is defined at synchronous

mechanical speed; the exact differential equation would be nonlinear

© Alberto Berizzi - Dipartimento di Energia

The mechanical swing equation cont’d 21

• Usually, H (Inertia constant ) is defined (2-9 MJ/MVA), as the ratio

between the stored kinetic energy in MJ at synchronous speed and
the machine rating in MVA: its range is not as wide as the range of M
 H=3 s is such that the kinetic energy stored can feed the rated
load for 3 s

• This relationship is in absolute values; by dividing by S, on the RHS

we have pu power, while speed and angles are in mechanical rad/s
and mechanical rad, respectively
• In Europe, the mechanical time constant T=2H is often used
© Alberto Berizzi - Dipartimento di Energia
Swing equation in pu 22

• Therefore, provided that angular speed and angles are in consistent

units, we can switch both to electrical angles (and pu power); they are
related by the pole pairs p:
  p m   p m
2H d  2
 pa  pm  pe
s dt 2

H d 2

 pa  pm  pe
 f dt 2

• Sometimes, the swing equation is also conveniently expressed as:

 d
 dt    s  s  pu  1

 d   s  p  p   1  p  p 
 dt 2 H m e
M
m e

© Alberto Berizzi - Dipartimento di Energia

Synchronous generator modelling 23

• In large perturbation angle stability studies, the transient state is

considered: therefore, each generator is represented by its internal
voltage E’ in series with the transient reactance X’
• The model is equal to the model for synchronous steady-state, except
for the numerical parameters

© Alberto Berizzi - Dipartimento di Energia

Possible simplifications: coherent machines 24

• In case of coherent machines, for example when they are in the

same power station and far from the fault, angles variations can be
considered the same and it is possible to obtain an equivalent
machine characterized (in the same pu base) by summing the
relevant swing equations, thus reducing the number of equations to
be solved (1=2)

2 H1 d 21
2
 pa1  pm1  pe1
s dt
2 H 2 d 2 2
2
 pa 2  pm 2  pe 2
s dt

2 H tot d 2
2
 patot  pmtot  petot
s dt
© Alberto Berizzi - Dipartimento di Energia
Possible simplifications: not coherent 25
machines

• In case of machines not coherent, it is possible to reduce the number

of equations by pairs of machines, to obtain an equivalent machine

d 21 s  pm1  pe1 

2
  
dt 2  H1 
d 2 2 s  pm 2  pe 2 
2
  
dt 2  H2 

d 21 d 2 2 s  pm1  pe1 pm 2  pe 2 
2
 2
   
dt dt 2  H1 H2 
2  H1 H 2  d 2
1   2  H 2  pm1  pe1  H1  pm 2  pe 2 
   
s  H1  H 2  dt 2
H1  H 2 H1  H 2
© Alberto Berizzi - Dipartimento di Energia
Possible simplifications: not coherent 26
machines

• Therefore, for each pair of non coherent machines, it is possible to

obtain an equivalent machine characterized (in absolute values) by
H12= H1H2/(H1+H2)
and by an accelerating power:
Pa12=(Pa1H2-Pa2 H1)/(H1+H2)
• If only a couple of non coherent machines is present, it is worth
noticing that, by arranging conveniently equations, the following can
be obtained:
2 H12 d 12 2
 pa12
s dt 2

• This proves that in a two-machine system, stability depends on the

difference of angles, not on their absolute values: if this difference is
decreasing in time, the system will be stable

© Alberto Berizzi - Dipartimento di Energia

Synchronous generator modelling: 27
transient state

• The Synchronous (lossless) machine model gives, in transient state

(not the same as at steady-state!):

𝑉𝐸′ 𝑉 1 1
𝑃 sin 𝛿′ sin 2𝛿′
𝑋′ 2 𝑋′ 𝑋′
𝑉𝐸′ cos 𝛿′ sin 𝛿′
• For anisotropic machines 𝑄 cos 𝛿′ 𝑉
𝑋′ 𝑋′ 𝑋′

𝑉𝐸′
𝑃 sin 𝛿′
𝑋′
• For isotropic machines 𝑉
𝑄 𝐸′ cos 𝛿′ 𝑉
𝑋′

© Alberto Berizzi - Dipartimento di Energia

 Power curve of a salient pole generator 28

• The curve (at no load under the

assumptions v=e=1) allows us to
determine the pre-perturbation
conditions
• Assuming v=1 pu, when the
generator is loaded on a
resistance at the pu power 0.6,
the excitation system must bring
e=1.164 pu
• In these conditions, =30° (0.54
rad)
• What would happen for =90°?

© Alberto Berizzi - Dipartimento di Energia

State transitions 29

• Let us consider a fault on the grid, close to the generator :

 State 0: initial state of the system, before the fault occurs. Steady-
state algebraic equations are to be considered (PF, possibly
integrated by controls and internal machine variables)
 State 1: immediately after the fault, the machine starts
accelerating (transient state equations) until the protection system
in the grid identifies the fault and trips the breaker
 State 2: the fault is cleared and hopefully a new steady-state will
be reached, that is, constant speed

• Simplifying assumptions:
 The power supplied by the turbine to the synchronous machine is
kept constant (the governor time constants are higher)
 The AVR effect is neglected, i.e., E’ is kept constant

© Alberto Berizzi - Dipartimento di Energia

Changes in the power curve 30

• When a fault modifies voltages

and/or impedances as seen by
the generator, the P-’ curve
changes
• The phenomenon can result in
loss of synchronism, if the
accelerating power is large
enough
• The turbine power is assumed
constant
• Initially, that power given by the
turbine determines a steady-
state value of ’, ’

© Alberto Berizzi - Dipartimento di Energia

Reactances during the fault 31

• The reactance as seen from

the generator changes during
the fault (from (b) to (c)), thus
modifying the power-angle
curve of the generator
• The reactance xF depends
on the type of fault (three-
phase, SLG, DLG, etc.)
 In case of three-phase solid fault,
it is zero
• The power-angle curve
changes depending on x’dF
(infinite, in case of solid 3-
fault)
• When the fault is cleared, the
power system is again
represented by circuit (b)
© Alberto Berizzi - Dipartimento di Energia
 Three-phase solid fault 32

• When a solid fault takes place at

generator terminals F, Pe goes to 0,
until the fault is cleared, at time tc
• During this time interval, ’ increases
• The difference in power is stored in
the machine as kinetic energy

𝑑 𝛿 𝜔
𝑝 0
𝑑𝑡 2𝐻
𝑑𝛿 𝜔 𝜔
𝑝 𝑑𝑡 𝑝 𝑡
𝑑𝑡 2𝐻 2𝐻 constant

𝜔 𝜔 𝑡
𝛿 𝛿 𝑝 𝑡 𝑑𝑡 ⇒ 𝛿 𝛿 𝑝
2𝐻 2𝐻 2

© Alberto Berizzi - Dipartimento di Energia

 Fault clearing: stable case 33

• 𝛿 increases in time as a parabula (d)

• When the fault is cleared, suddenly pe
increases to the corresponding pe(𝛿 ), point
5, larger than pm
• The accelerating power is now negative, the
rotor slows down, but the angle 𝛿 keeps
increasing until the speed is back to the
synchronous speed, point 6, and
corresponding 𝛿
• Then, the speed keeps decreasing, getting
lower than the synchronous speed
• At this point, as the speed is lower than
synchronous speed, 𝛿 also decreases, etc.
• In the absence of damping, this is a never-
ending phenomenon (electromechanical
oscillations), limit cycle
Machowski, Bialek, Bumby:
• With damping, synchronism is not lost Power system Dynamics
© Alberto Berizzi - Dipartimento di Energia
Fault clearing: instable case 34

• In this case, the decelerating

power between 4 and 8 is not
sufficient to limit 𝛿
• When point 8 is overtaken,
power is accelerating again
• 𝛿 still increases and the
machine loses the step, resulting
in an asynchronous, and
accelerated, rotation

© Alberto Berizzi - Dipartimento di Energia

Equal-area criterion 35

• It is a criterion based on energy, not popular for multimachine systems

• Let’s define the relative speed with respect to synchronous speed:

• By substituting for r, we get:

2 H d r
 pm  pe
s dt
• And multiplying both sides by r
2𝐻 𝑑𝜔 𝑑𝛿 𝐻𝑑 𝜔 𝑑𝛿
𝜔 𝑝 𝑝 ⇒ 𝑝 𝑝
𝜔 𝑑𝑡 𝑑𝑡 𝜔 𝑑𝑡 𝑑𝑡

© Alberto Berizzi - Dipartimento di Energia

 Equal-area criterion cont’d 36

• We can integrate this expression

𝐻𝑑 𝜔 𝑑𝛿
𝑝 𝑝
𝜔 𝑑𝑡 𝑑𝑡
𝐻
𝑑 𝜔 𝑝 𝑝 𝑑𝛿
𝜔

• This is a general form: we can use it for any 1 and 2. Consider points
0 and e. Both are characterized by the synchronous speed, i.e., r=0:

𝑝 𝑝 𝑑𝛿 0 𝑝 𝑝 𝑑𝛿 𝑝 𝑝 𝑑𝛿

⇒ 𝑝 𝑝 𝑑𝛿 𝑝 𝑝 𝑑𝛿

© Alberto Berizzi - Dipartimento di Energia

 Equal-area criterion cont’d 37

• The latter equation represents the equal-area criterion, given by the

integrals at both sides of the equal sign:
 If ’e reaches ’max, then
instability will occur: loss of
synchronism
 Otherwise, oscillations will show
up, as described
In that case, the issue is getting a
sufficient damping

© Alberto Berizzi - Dipartimento di Energia

Physical meaning of the criterion 38

• Considering the expression

𝐻
𝑑 𝜔 𝑝 𝑝 𝑑𝛿
𝜔

 When pm > pe, the area below pm is proportional to an increase of

kinetic energy
 Viceversa, for pe > pm, the area above pm is proportional to a
decrease of kinetic energy,
• The critical clearing time (and the corresponding critical angle) is the
maximum clearing time that guarantees that a stable situation is
recovered.
• When the breaker clears the fault exactly at the critical clearing time,
exactly the angle ’max will be reached.
• Hence, for ’>’max, synchronism is lost.
• ’max depends on the initial conditions, i.e., on pm, and on the clearing
time
© Alberto Berizzi - Dipartimento di Energia
Computation of critical quantities for three- 39
phase solid fault

• Compute the two areas involved (pe=0 during the fault):

𝐴 𝑝 𝑝 𝑑𝛿 𝑝 𝑑𝛿 𝑝 (𝛿 - 𝛿 )

𝐴 𝑝 sin 𝛿 𝑝 𝑑𝛿 𝑝 (cos 𝛿 -cos 𝛿 )-𝑝 (𝛿 -𝛿 )

• As they must be equal, at critical time:

( - )= ( - )- ( - )

( - )= ( - )

= -

© Alberto Berizzi - Dipartimento di Energia

 Computation of critical time and angle 40

= -
=
= -
= − −
- −

And from the relationship

• tcr decreases (worse and worse)

• when inertia decreases
4𝐻 𝛿 𝛿 • when initial load increases
𝑡
𝜔𝑝

© Alberto Berizzi - Dipartimento di Energia

Stability margin 41

• Area 6-7-8 divided by area 4-5-

6-8 is a Transient stability
margin index
• Area 1-2-3-4 is proportional to ,
(base of the rectangle) i.e., to
the square of the time

𝜔 𝑡
𝛿 𝛿 𝑝
2𝐻 2

• Hence, it is very critical to

increase the clearing time!

© Alberto Berizzi - Dipartimento di Energia

Pre-fault load influence 42

• If the pre-fault load increases, for a three-phase fault,

 the accelerating area increases proportionally
 the maximum angle ’ increases proportionally (look at the
expression of ’as a function of pm)
• The critical clearing time decreases 𝜔 𝑡
𝛿 𝛿 𝑝
2𝐻 2

© Alberto Berizzi - Dipartimento di Energia

Example on criticality of the fault duration 43

• Example 11.4 from Sarma’s book. A fault is self extinguished

at text= 3 cycles= 0.05s
 ’ 0=23.95°,
 Fault cleared at ’ d=28.44°,
 Maximum angle reached ’ e=40.12°,
 ’ max=(180°-23.95°)=156.05°stable
• By the equal area criterion (60 Hz)
 ’ crit=88°>’ d=28.44°stable
 tcrit=0.189s, i.e., 11.38 cycles
• Simulation with D=0: increase t and check ’ crit
• Case with D=1 and damping appears (try tcr=0.19s)
• Case with PG=150 MW and check that ’ max decreases (about 180-
37=143°), about tcrit=0.1152

© Alberto Berizzi - Dipartimento di Energia

Effect of the distance of fault 44

• Stability depends on the equivalent reactance as seen from the

generator:
 The greater the distance, the closer the post-fault power-angle
curve to the pre-fault power angle curve. Hence, less severe the
fault and larger the critical clearing time

© Alberto Berizzi - Dipartimento di Energia

 An example with change of post-fault 45
impedance

2.5
Before the fault
• Let us assume that the fault occurs During the fault
Post fault clearing

along a line: 2

 initially, the equivalent power curve 1.5

is subject to the decrease of the

P [p.u.]
voltage (increase of equivalent 1

impedance) 0.5

 then, when the fault is cleared

(breakers open), the equivalent 0
0 0.5 1 1.5
delta [rad]
2 2.5 3 3.5

impedance is higher than the initial

equivalent impedance

© Alberto Berizzi - Dipartimento di Energia

Short-circuit with re-closure of the breaker 46

• Opening and closing the breaker

modifies impedances
• We assume the fault at a given
distance
• If the re-closure succeeds, the
final equivalent circuit is (b)
again, otherwise, a further
tripping is necessary, and the
final configuration is (d)
• We must draw three P-’ curves

© Alberto Berizzi - Dipartimento di Energia

 Short-circuit with re-closure of the breaker 47
on a temporary fault

• Case (a) represents a stable

case: re-closure on a
vanished fault: maximum
angle is point 9 and
equilibrium point is point 1
• Case (b) is an unstable case:
re-closure occurs too late and
even though the fault is
vanished, synchronism is lost
• Positive factors of re-closure:
 The maximum angle moves
on curve a, i.e., on the right
(point 9)
 Decelerating area increases

© Alberto Berizzi - Dipartimento di Energia

 Short-circuit with re-closure of the breaker 48
on a permanent fault

• If the fault is permanent, after re-

closure a final tripping of the
breaker is necessary
• In the stable case (b), the initial
load is reduced, and the same
tripping times of the breaker
allow the final operation in the
equilibrium point 1’, different
from the initial equilibrium point

© Alberto Berizzi - Dipartimento di Energia

 Power oscillations 49

• Depending on the initial loading

Pm, p(t) shows different shapes
• Considering the case without
re-closure: if ’ overcomes /2,
two hunchbacks can be seen
• If the stability margin is large,
they do not turn up

© Alberto Berizzi - Dipartimento di Energia

Example on CCT evaluation 50

• Example 11.6 from Sarma’s book. A three-phase solid fault occurs on

line 1-3 at bus 3 and it is cleared by opening the two CBs.
• By the equal area criterion (60 Hz)
 ’ cr=111.5°, tcrit=0.34s
• Show shape of real power
• Case with PG=150 MW and check that ’ crit decreases

© Alberto Berizzi - Dipartimento di Energia

Effect of change in flux 51

• So far, we have assumed that flux in the synchronous generator is

kept constant: E’=const.
• Actually, in rotor circuits, some energy is lost, and this results in a
decrease of E’ in time, reducing margins and lowering the power
curve
• Therefore, adopting E’ constant may give non-conservative results
• If the clearing time is short, this effect is negligible

© Alberto Berizzi - Dipartimento di Energia

 Effect of the AVR 52

• A strong action of the AVR may

avoid loss of synchronism
• Let us consider the case of a fault
at a given distance
• Case (a) is without the action of
AVR
• Case (b) takes into account the
increase of E’ due to the AVR
 The accelerating area decreases
 The decelerating area increases
• A possible drawback is that this
can result in the increase of the
second and further oscillations
• After point 6 the AVR should
decrease the excitation,
independent of the terminal
generator voltage

© Alberto Berizzi - Dipartimento di Energia

Importance of AVRs for rotor angle stability 53

• Large-disturbance performance of a multi-machine power system

depends on adequate synchronizing power flows between
synchronous generators to prevent loss of synchronism of any
generator on the system
• High gain excitation systems are employed to increase synchronizing
power flows and torques
• The decay of oscillations, following for example the initial transient
(usually the first swing) is dependent on the presence of damping
torques acting on the generator rotors
• Damping torques may be degraded significantly by high gain
excitation systems such that, if the net damping torque is negative,
instability occurs
• To counter this type of instability, positive damping torques can be
induced on generators by installing continuously-acting controllers
known as Power System Stabilizers

© Alberto Berizzi - Dipartimento di Energia

Some examples of the importance of fault 54
location

© Alberto Berizzi - Dipartimento di Energia

Example on the IV order machine 55

• Example 11-10 from Sarma’s book

• Determine the fault clearing time (about 0.3-0.31s)

• Observe that the system is naturally damped by the model of the field
(D=0)
• Add plot of power

© Alberto Berizzi - Dipartimento di Energia

 Unsymmetrical faults 56

• The severity depends on the power

that can be transferred to the system
during the fault, and before its clearing
• In case of non symmetrical fault, a
share of the mechanical power can be
transferred to the network
• In (b) the SLG case is considered.
Acceleration is given by area 1-9-10-
12: the increase of is lower (with the
same clearing time)
• After fault clearing, the operating point
is 11
• The stability margin is a great deal
larger than in the three-phase case
(dashed)

© Alberto Berizzi - Dipartimento di Energia

Other key factors 57

• In modern power systems, stability is a more and more important

problem:
 generator sizes are ever increasing, but rotor masses do not
increase at the same rate, due to more efficient cooling systems:
the H constant keeps decreasing
 Up-rating processes tend to increase generator X’
 RES based generation is typically inertialess

© Alberto Berizzi - Dipartimento di Energia

How to increase stability? 58

• According to the equal-area criterion, the following comments can be

made:
 Reduction of clearing time, thus reducing the accelerating energy,
when voltages are particularly low
 Reducing re-closing time, if re-closure is implemented
 Increasing generator inertia (e.g., flywheels)
 Single-pole operation of breakers can increase critical clearing
time by 2 to 5 cycles
 Quick excitation systems, able to reduce the first rotor angle swing
(within 0.5-1.5 cycles)
 Fast-valving: it allows a sudden reduction of steam admission into
steam turbines turbine (it can last about 1 s), thus reducing pm
 Braking resistors, to be inserted when the breaker is open, to
avoid any reduction of the generator output
 Insertion of series capacitors to be connected on unfaulted lines
after fault clearing, to reduce the equivalent impedance
© Alberto Berizzi - Dipartimento di Energia