POWER SYSTEM STABILITY

The electric power system is a dynamic and

highly nonlinear system. The dynamics are
due to changes such as
(a) Load and generation changes
(b) Switching, e.g. circuit breaker operations
(c) Faults such as short circuits
(d) Lightning surges, etc.
The dynamics can be classified by the speed of
occurrence. For example

Phenomenon Speed (sec.)

1 Lightning surges 10-9 → 10-3
2 Switching surges 10-9 → 10-3
3 Harmonics 10-3 → 10-2
4 Subsynchronous resonance 10-2 → 1.0
5 Transient stability 10-2 → 1.0
6 Dynamic stability 10-1 → 10.0
7 Load frequency control 10-1 → 102
8 Turbine and boiler control 10-1 → 103

The models needed to study these dynamics

vary in detail depending on the speed of
occurrence. For example,

1
9 Lightning surges employ distributed
parameter models (partial differential
equations)
9 Slower phenomena such as gradual load
and generation changes are modeled by
lumped parameters (algebraic equations)

Definitions

Stability
Power system stability refers to the ability of
the various machines in the system to remain
in synchronism (stay in step), with each other
following a disturbance.
Stability studies may be classified upon the
nature of the disturbance as follows.

Steady state stability

which refers to the ability of the various
machines in the system to remain in
synchronism after a small and slow
disturbance, such as a gradual change in load.
In this type of stability study, a very simple
synchronous machine model is used (e.g., the
generator is treated as a constant voltage

2
source). In addition, this kind of stability
studies is usually analyzed by load flow to
check phase angles and bus voltages.

Dynamic stability
which refers to the ability of the various
machines in the system to remain in
synchronism after a small disturbance, such as
badly set automatic controls.
In this type of stability study, the excitation
system and turbine-governing system are
represented along with synchronous machine
models by linear differential equations.

The steady state and dynamic stability studies

are less extensive in scope and involve one or
just a few machines undergoing slow changes
in load.
The nonlinear differential and algebraic
equations of both study cases can be replaced
by a set of linear equations and then solved by
linear analysis methods. The solution
technique is to examine the stability of the
system under incremental variations about the
equilibrium point.

3
Transient stability
which refers to the ability of the various
machines in the system to remain in
synchronism after a sudden large disturbance,
such as a sudden change in load, a fault, loss
of a generator, and a switching operation.
The nonlinear differential and algebraic
equations do not allow the linearization
process to be used, and must be solved by
direct methods or by iterative step-by-step
procedures.
Transient stability problems can be subdivided
into first-swing (which is based on a simple
generator model) and multi-swing (which must
consider effects of generator control systems
along with generator model) stability
problems.
The period under investigation can vary from a
fraction of a second (for first-swing case) to
over ten seconds (for multi-swing case).
In all stability studies, the objective is to
determine whether or not the rotors of the
machines being perturbed return to constant
speed operation.

4
The Swing Equation

This is an equation relating the rotational

masses to the generator electrical parameters.

J
Te
Tm d 2θ m
J = Ta = Tm − Te N-m
dt 2
where
J → total moment of inertia of the rotor
masses
θm → the angular displacement of rotor with
respect to a stationary axis in
mechanical radians
t → time in seconds
Tm → mechanical or shaft torque supplied by
prime mover less rotational losses
Te → electrical or electromagnetic torque
Ta → net accelerating torque
Ta = 0 → constant frequency (60 Hz)
Ta > 0 → net accelerating torque.
Speed will increase indefinitely
Ta < 0 → net decelerating torque.
Speed will decrease indefinitely

5
Define θm = ωsm t + δm
Where
ωsm → synchronous speed of the machine in
mechanical radians/sec
δm → the angular displacement of rotor from
synchronously rotating reference in
mechanical radians
Hence, the rotor angular velocity is
dθ m dδ
= ω sm + m
dt dt
and, the rotor angular acceleration is
d 2θ m d 2δ m
2
=
dt dt 2
Equation of motion becomes
d 2δ m
J 2
= Ta = Tm − Te
dt
define, ωm = dθm / dt then,
d 2δ m
Jω m = Pa = Pm − Pe
dt 2
Jωm = the angular momentum of the rotor
Define, M = Jωm is the inertia constant of the
machine. Then approximately,
d 2δ m
M 2
= Pa = Pm − Pe
dt

6
i.e. we assume machine rotor speed does not
vary greatly from synchronous speed.
To account for frictional torques or eddy
currents losses, a deceleration torque is
introduced.
d 2δ m dδ m
M = Pm − Pe − Da
dt 2 dt
Define,
1
Stored kinetic energy at synchronous speed 2 J ω 2
sm
H= =
Generator rating S mach
1
Mω sm
H=2 MJ/MVA
S mach
H is called H-constant of the machine. The
unit of H is sec. H is preferred over M because
it falls within a fairly narrow range (1-10).
Thus,
dδ m
Pm − Pe − Da
2H d δ m
2
Pa dt
= =
ω sm dt 2 S mach S mach
The electrical angles relates mechanical angle
by
p
δ (t ) = δ m (t )
2
where p is the number of poles
7
In terms of electrical angles,
2 H d 2δ dδ
= P = P − P − D pu
ω s dt 2
a m e
dt
This equation is called swing equation. Here
δ is in electrical degrees
ωs is electrical synchronous speed radians/sec
The swing equation is usually expressed in
two first order forms as
2 H dω dδ
= Pm − Pe − D
ω s dt dt
dδ
= ω − ωs
dt
Comments
1. The plot of δ versus time yields the swing
curve. The stability of a power system can
be determined from the plot of the swing
curves of all machines in the system.
2. The mechanical input power Pm in the
swing equation is considered constant
because conditions in the electrical
network can be expected to change before
the control governor can cause the turbine
to react.
3. If Pa = 0, i.e Pe = Pm the machine operates
at steady state synchronous speed

8
4. Changes in Pe could result from severe
load changes, network faults ,…etc.
5. The solution of the swing equations is very
difficult even in case of single-machine-
infinite-bus system. Such difficulties arise
due to the nonlinear nature of the equations

Synchronous Machine Model (Simplified)

E'
jIxd'
I jxd' +
δ Vt
Vt
E '∠δ
_ Reference
I
E ' → transient internal voltage
δ → machine angle relative to synchronously
rotating reference

Stability Criterion

Case A: Machine connected to an infinite

bus

jxd' I jxl
E ∠δ'
V ∠θ '

9
Total reactance, X = x d' + x l and
E 'V
Pe = sin θ
X
where θ = δ − θ '
• ••
For infinite bus, H = ∞ and θ ' = ωs , θ ' = 0.
•• ••
Then θ = δ .
Assuming D = 0, machine dynamics will be
governed by
•• ωs ⎡ E 'V ⎤
θ= P
⎢ m − sin θ ⎥
2H ⎣ X ⎦
Pe vs θ can be plotted as shown
Pe

A B Pm

θ1 θ2 θ

A and B are equilibrium points. Let us consider

slight variations about equilibrium points.
Point A: If θ > θ1 or θ < θ1, machine will slow
down or speed up to restore θ. Point
A is a stable equilibrium point.
Point B: If θ > θ2 then Pm > Pe and machine
will speed up and
10
 9 Speed will increase dangerously
9 Synchronous operation cannot be
maintained
9 Machine loses synchronism or pulls
out of step
Point B is an unstable equilibrium
point.

Case B: Machine connected to another

machine

jxd' 1 jxl jxd' 2

E1' ∠δ 1 E2' ∠δ 2
•• •• ••
Define θ = δ1 − δ 2 then θ = δ1 − δ 2 . It can be shown
•• ωs
δ1 = [Pm1 − Pe1 ]
2 H1
•• ωs
δ2 = [Pm 2 − Pe 2 ]
2H 2
For a 2 machine system,
Pe1 = -Pe2 = Pe (say)
Pm1 = -Pm2 = Pm (say)
Define δ12 = δ1 − δ 2 , then
•• ωs
δ12 = [Pm − Pe ]
2 H eq

11
where
1 1 1
= +
H eq H1 H 2
Here, stability is governed by how one
machine pulls apart (δ1- δ2)from the other.

Case C: Multi-machine system

In this case,
2Hi ••
δ i = Pmi − Pei , i = 1,2,..., n
ωs
each δi is measured w.r.t. synchronous
reference. To determine stability we need a
reference to observe the pulling apart of the
machines. Commonly used references
1. Make the angle of the largest machine in
the system the reference angle and plot
(δi- δlargest machine)
2. Define a Center Of Angle (COA)
1
COA = ∑
number of machines i
δ i (t )

and plot δi(t) vs COA for all machines

12
Power-Angle Equation

For a given network configuration with

constant voltage magnitudes of the transient
internal voltage of two machines (or a machine
and an infinite bus) E1' and E2' , and for
negligible armature resistance, the power-
angle equation of the network is defined as
Pe = Pmax sin δ
where
E1' E2'
Pmax = = E1' E2' Y12
X 12
and X12 is the transfer reactance between E1'
and E2' , Y12 is the transfer admittance. Note
that E1' , E2' , Y12, and X12 are all magnitude
values.
It can be observed from the power-angle
equation and the power-angle curve that the
steady state stability limit for synchronous
machine connected to an infinite bus is when δ
≤ 90˚ or dP/dδ > 0.
The possible power-angle swing curves for the
stability process is shown below.

13
Example 1

Consider the one-line diagram and the

corresponding reactance diagram of the two-
machine system shown below.

Let X s = X s = 0.4 pu, X T = X T = 0.1 pu,

G M 1 2

and X l =1.2 pu
a) Analyze whether the system is stable
enough in the steady state to deliver 1.2
pu power at the motor terminals with
rated terminal voltage and 1.0 pf

14
 b) Repeat (a) for a second line of the same
size in parallel with the original.
c) What is the maximum power that can be
delivered in part (b).

Solution:

a) For VM = 1.0∠0 D pu, I = 1.2∠ 0D pu

E M = VM − jIX sM = 1 − j 0.48 = 1.109∠ − 25.6 D pu
EG = VM + jI ( X T2 + X l + X T1 + X sG ) = 1 + j 2.16 = 2.38∠65.2D pu
δ GM = δ G − δ M = 90.8D
The system is unstable

b) In this case, X l =0.6 pu

EM = 1.109∠ − 25.6D pu
EG = VM + jI ( X T2 + X l + X T1 + X sG ) = 1 + j1.44 = 1.753∠55.2D pu
δ GM = δ G − δ M = 80.8D
The system is stable

c) The transfer reactance between EG and EM

is
X = X sM + X T2 + X l + X T1 + X sG = 1.6 pu
1.109 × 1.753
Pmax = = 1.215
1 .6

15
Example 2

In the system shown, the machine is delivering

power of 1.0 pu, and both the terminal voltage
and the infinite-bus voltage are 1.0 pu.
Assume a three-phase to ground fault at point
P. The fault is cleared by opening the circuit
breakers at each end of the faulted line.
Determine the power-angle equation for the
following cases:
a) Pre-fault b) During fault c) After fault

Solution:
Choose the infinite bus voltage as the
reference.
a) Pre-fault power-angle equation

16
The transfer reactance between the Vt and the
infinite bus is 0.1 + 0.4/2 = 0.3 pu. Assume α
is the angle of Vt. Hence,
VtV∞ 1× 1
Pe = sin α = sin α = 1
X 0.3
This gives α = sin-1 0.3 = 17.458˚.
Therefore, Vt = 1.0∠17.458 = 0.954 + j 0.300
D
pu
Vt − V∞ 1.0∠17.458D − 1.0∠0D
I= = = 1.0 + j 0.1535 = 1.012∠8.729D pu
jX j 0.3
E ' = Vt + jIxd' = 1.0∠17.458 D + j 0.3 × (1.0 + j 0.1535) = 1.05∠28.44 D pu
The total reactance = 0.2 + 0.1 + 0.4/2 = 0.5 pu
Hence,
1.05 × 1
Pe = sin δ = 2.10 sin δ pu
0 .5
where δ is the machine rotor angle with
respect to the infinite bus.
Note that under steady state condition, the
rotor angle δ of the machine is equal to 28.44˚,
in which Pe = Pm = 1.0 pu.

b) During-fault power-angle equation

17
The values shown are admittances in per unit.
The effect of the fault is clearly shown by
redrawing the reactance diagram as

Under the assumption of constant flux linkages

in the machine, the transient internal voltage of
the machine remains at E '
= 1 . 05∠ 28. 44 D
pu
To find the transfer admittance between E '
and the infinite bus voltage, establish the
admittance matrix as
⎡− 3.333 0 3.333 ⎤
Ybus = j ⎢⎢ 0 − 7.5 2.5 ⎥⎥
⎢⎣ 3.333 2.5 − 10.833⎥⎦

Since bus 3 has no external source connection,

it may be removed by the node elimination
method. The reduced admittance matrix is
⎡Y11 Y12 ⎤ ⎡− 2.308 0.769 ⎤
⎢Y Y ⎥ = j ⎢ 0.769 − 6.923⎥
⎣ 21 22 ⎦ ⎣ ⎦
Hence, the desired power-angle equation, is
Pe = E 'V∞Y12 sin δ = 1.05 × 1 × 0.769 sin δ = 0.808 sin δ pu

18
c) After-fault power-angle equation

The total transfer reactance = 0.7 pu

Hence,
1.05 × 1
Pe = sin δ = 1.5 sin δ pu
0 .7
The power-angle curves of all cases are shown
below.

19
Comments
1. Pm is constant in all cases
2. Pm intersects the steady state case (a) at
28.44˚, i.e. at this operating point Pa = 0
3. Pa during-fault & after-fault depends upon
the rotor angle at that time.
4. It can also be seen from the power-angle
curves that Pm = Pe at 151.56˚. However,
this operating point is not acceptable
because it has a negative slope (unstable
equilibrium point).
5. The swing equations in all cases can be
written as follows.
••ω
(a) Pre-fault δ= [1.0 − 2.100 sin δ ]
s
2H
•• ωs
(b) During-fault δ= [1.0 − 0.808 sin δ ]
2H
•• ωs
(c) After-fault δ= [1.0 − 1.500 sin δ ]
2H
6. The swing equations are highly nonlinear
and can only be solved by
(a) Graphical solution (equal area
criterion) for one- or two-machine
systems
(b) Numerical integration methods such as
trapezoidal, Runge-Kutta, …etc.

20