Stability 4

1.0 Introduction
In the previous notes (Stability 3), we developed
the equal area criterion, which says that
For stability, A1=A2, which means the
decelerating energy (A2) must equal the
accelerating energy (A1) in order for the
system response to be stable.
Analytically, we have that
clear

max

clear

0
P
M Pfaultd

0
P

P
fault M d

A1

A2

Figure 1 below illustrates a stable case.

Pe

2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0

Ppre
(g) Ppost

(f)
(e)

(a)

A2

(d)

A1
Pfault

(c)
(b)

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 a

0
clear
max

Fig. 1
1

(1)

Figure 2 below illustrates an unstable case.

Pe

2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0

Ppre
Ppost

A2

A1
Pfault

(c)
(b)

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 a

0
clear
max

Fig. 2
Everything about Figs. 1 and 2 are the same
with one exception, the clearing angle (clear) in
Fig. 2 is greater than the clearing angle in Fig. 1.
In other words, Fig. 2 assumes that the speed of
the protection system is slower than the speed of
the protection system in Fig. 1.
In these notes, we want to develop expressions
for computing critical clearing angle.

2.0 Critical clearing angle

The critical clearing angle will occur when the
equal-area criterion is satisfied and the
maximum angle is max=180-0. Such a case is
illustrated in Fig. 3.
Pe

2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0

Ppre
Ppost

A2

A1

Pfault

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 a

0
clear
max

Fig. 3
We want to compute the critical clearing time.
We will denote it as clear=cr.

To do this, lets define the following:

Ppre Ppre max sin

(2)

Pfault Pfaultmax sin r1Ppremax sin

(3)

Ppost Ppostmax sin r2 Ppremax sin

(4)

where
Ppremax, Pfaultmax, and Ppostmax are the amplitudes
of the power-angle curves for the pre-fault,
fault-on, and post-fault networks, respectively;
0<r1<1 where r1=0 corresponds to a threephase fault at the machine terminals, and r1=1
corresponds to no-fault at all.
0<r2<1 where r2=0 corresponds to a threephase fault at the machine terminals that is not
cleared, and r2=1 corresponds to a temporary
fault (fault is removed without protective relay
action to also remove a circuit and weaken the
transmission)
Lets first compute A1.

A1

clea r

0
P
M Pfaultd

clea r

0
(
P
M Pfaultmax sin )d

P Pfaultmax cos
0
M

clea r
0

(5)

PM0 ( clear 0 ) Pfaultmax (cos clear cos 0 )

Now lets compute A2.

A2

max

0
P

P
post M d

clear

max

0
P
sin

P
M d
postmax

clear

Ppostmax cos P
0
M

max
clear

(6)

Ppostmax (cos clear cos max ) PM0 ( clear max )

If the system is stable, then A1=A2. So lets

equate the expressions in eq. (5) and (6), below.

A1 PM0 ( clear 0 ) Pfaultmax (cos clear cos 0 )

Ppostmax (cos clear cos max ) PM0 ( clear max ) A2 (7)

Expand:
PM0 clear PM0 0 Pfault max cos clear Pfault max cos 0
Ppostmax cos clear Ppostmax cos max PM0 clear PM0 max (8)
0
Notice there is a PM cr on both sides, and so:

PM0 0 Pfault max cos clear Pfault max cos 0

Ppostmax cos clear Ppostmax cos max PM0 max

(9)

Lets put all terms with cosclear on the left side

and everything else on the right:
Pfault max cos clear Ppostmax cos clear

Pfault max cos 0 Ppostmax cos max PM0 max PM0 0 (10)

Factor out the cosclear term on the left and the

PM0 term on the right:
cos clear ( Pfault max Ppostmax )

Pfault max cos 0 Ppostmax cos max PM0 ( 0 max ) (11)

Divide by the term in parentheses on the left:

cos clear

Pfaultmax cos 0 Ppostmax cos max PM0 ( 0 max )

( Pfaultmax Ppostmax )
6

(12)

Now eq. (12) is true as long as the system

response is stable (if it is not stable, then the
equal-area criterion is not satisfied and therefore
eq. (7) is invalid).
But if the system response is marginally stable,
then the clearing angle will be the maximum
possible angle for which we can clear and still
retain stability, i.e., it is the critical clearing
angle, and so in this case, clear=cr. In addition,
the maximum angle must be the unstable
equilibrium, which is max=180-0. Making these
substitutions into eq. (12) results in
cos cr

Pfault max cos 0 Ppostmax cos( 0 ) PM0 ( 0 0 )

( Pfault max Ppostmax )

(13)
Recalling that cos(-x)=-cos(x), and noting on
the right-hand-side that we can combine the two
0 terms inside the brackets, we get:

cos cr

Pfault max cos 0 Ppostmax cos 0 PM0 (2 0 )

(14)

( Pfault max Ppostmax )

Now recall eqs. (3) and (4), which imply that:

Pfaultmax r1 Ppremax

(15)

Ppostmax r2 Ppremax

(16)
Substituting eqs. (15) and (16) into (14), we get:
cos cr

r1Ppremax cos 0 r2 Ppremax cos 0 PM0 (2 0 )

(r1Ppremax r2 Ppremax )

(17)

Factoring out the Ppremax from the bottom and

dividing it through all terms in the top, and
rearranging, results in
r1 cos 0 r2 cos 0
cos cr

PM0
Ppremax

(r1 r2 )

Lets consider a few cases:

(2 0 )

(18)

Case 1, Temporary fault at machine terminals:

The fact that it is a temporary fault means that
the post-disturbance network is the same as the
pre-disturbance network, therefore r2=1.
The fact that it is a three-phase fault at the
machine terminals means that the ability to
transmit power to the infinite bus, during the
fault-on period, is zero. Therefore r1=0.
Applying these values to eq. (18) results in:
cos 0
cos cr

PM0

PM0
Ppremax

(2 0 )

(19)

( 2 0 ) cos 0

Ppremax

But recall that

PM0 Ppremax sin 0

(20)
Substitution of (20) into (19) results in
cos cr

Ppremax sin 0
Ppremax

( 2 0 )

(21)

Or,

cos cr sin 0 ( 2 0 ) cos 0

(22)
The above is a closed form solution for the
critical clearing angle for the condition of a
temporary three phase fault at the machine
terminals.
Recall the example introduced in the notes
called Stability 2 for the below system:
j0.1

Xd=j0.2

j0.4

V= 1.0<0

j0.4
Bus 1

Bus 3

Bus 2

|Vt |= |V1 |=1.0

Fig. 4
In those notes, we determined that the angle
between the generator internal voltage and the
infinite bus is a=28.44. This is 0, and it is for
the same system that is characterized in these
notes by Fig. 3. Using 0=28.44=0.4964 rad in
eq. (22) results in
10

cos cr sin 0 ( 2 0 ) cos 0

sin(0.4964)( 2(0.4964)) cos(0.4964)
0.4763( 0.998) 0.8793
1.021 0.8793 0.1417

Therefore we have that

cos cr 0.1417 cr cos 1 0.1417 1.4286rad
In degrees, this is 81.85. Reference to Fig. 5,
which is a hand-approximation for this case,
suggests this angle is quite reasonable.
Pe

2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0

Ppre
Ppost

A2

A1
Pfault

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 a

0
clear
max

Fig. 5

11

3.0 Critical clearing time

Lets consider our case of a temporary threephase fault at the machine terminals. To obtain
information on clearing time, we need to look at
the differential equation characterizing this
system, which is:

2 H
(t ) Pa , pu PM0 Pe

e 0

(23)

The right hand-side is of course 0 before the

fault (no acceleration), but just after the fault, in
this case, Pe goes instantly to 0. We therefore
have that

2 H
(t ) PM0

e 0

(24)

And so we see that just after the fault, there is

non-zero acceleration, but that acceleration is
constant since the right-hand-side is constant!
Equation (22) may be rewritten as

12

e 0 0

(t )
PM
2H

(25)

Rewrite the left-hand-side of eq. (25) as

d d e 0 0

(t ) 2

PM
dt
dt 2 H
2

(26)

Multiply both sides by dt:

e 0

PM0 dt

(27)
Now integrate on the left from (0)=0 (initial
state is zero velocity) to (t) and on the right
from t=0+ to t:

2H

(t )

e 0

d (t )
P dt

2H

( 0 ) 0

(t )

0
M

e 0

0
M

P t

(28)

(29)
Now express the left-hand-side as the derivative
of (t)

2H

13

d e 0 0

PM t
dt 2 H

(30)

Multiply both sides by dt:

e 0

PM0 tdt

(31)
Now integrate the left-hand-side from 0 to (t),
and the right-hand-side from t=0+ to t:
(t )

2H
t

e 0

d 2H P
0

0
M

tdt
(32)

(t ) 0

e 0

0 2
M

P t

(33)
Now recall we have the critical clearing angle
(t)=cr, and we are attempting to find the time
for which we reach this angle. So solve eq. (33)
for time to obtain:

4H

4H
t (t ) 0 0
PM e 0
14

(34)

When the angle is the critical clearing angle, we

obtain:

4H
tcr cr (t ) 0 0
PM e 0

(35)

So, lets compute the critical clearing time for

our machine. The only other thing we need to
know is the inertia constant H. We can assume
that it is H=3.0 sec on the machine base. With
PM0 1.0 , e0=377, 0=28.44=0.4964 rad, and
cr 1.4286rad , we have
4*3
tcr 1.4286 0.4964
1* 377
0.1723 sec
What if the inertia constant was 5? In this case,
we would obtain:
4*5
tcr 1.4286 0.4964
1* 377
0.2224 sec
The larger the machine, the longer it takes to
accelerate it.
15