Notes 20: Short circuit analysis

20.0 Introduction
The goal of short circuit analysis is to
identify the range of currents that will flow
in the network under faulted conditions.
The range of currents are identified by the
minimum and maximum currents that flow
under faulted conditions.
What do we mean by faulted conditions?
These are conditions that include one or
more short circuits somewhere in the
network.
Faulted conditions are highly undesirable
because the currents that flow under faulted
conditions are normally extremely high. If
uninterrupted, such currents cause very high
heating in the conductors and likely result in
damage to at least one of them.
1

Example:
Consider the very simple single phase
distribution system of Fig. 1. The per unit
impedance between the generator and bus 2
is 0.03125, between bus 2 and bus 3 is 0.10,
and between bus 3 and bus 4 is 0.025. The
load has a per unit impedance of 3.125. The
generator voltage is 1.0 per unit.
(a) Compute the normal load current in perunit.
(b)
Compute the current for a fault at bus 4
in per-unit.

Fig. 1
Solution:
(a)
The normal load current is easy to get:
1.0
I

0.3048 pu (1)
0.03125 0.1 0.025 3.125
(b) A fault at bus 4 will short the load.
Load

The resulting fault current will be:

1.0
I

6.4 pu
(2)
0.03125 0.1 0.025
fault

Note that in this example the fault current

magnitude is 21 times the normal load
current.
Another important observation that we can
make from this very simple example is that
the current into the fault entirely comes from
the generator.
There are two additional comments that we
can make from this observation.
Effect of load: First, this implies that no
fault current comes from the load. In fact,
the portion of the load downstream from the
fault has no influence on the fault current.
This makes complete sense if you examine
the faulted circuit, as shown in Fig. 2.

Fig. 2
To make the previous point very clear,
consider a fault at bus 3. Then the circuit is:

Fig. 3
and the fault current calculation is:
1.0
I

7.62 pu
(3)
0.13125
We note in the previous calculation that the
only part of the circuit that was of influence
was the part between the generator and the
fault. As a result, the 0.125 part of the
impedance is not included.
fault

Effect of source: The voltage at the source

and the impedance between the source and
the high side of the transformer always
influence the fault current for a fault located
anywhere on the feeder.
Of course, a distribution feeder is never
actually fed by a single generator, as we
have indicated in our example. Rather, a
distribution feeder is fed by the transmission
(or subtransmission) system.
Nonetheless, we may obtain an equivalent
source between the high side of the
transformer and the remainder of the
transmission network. We will discuss some
ways to do this a bit later. For now, you
should be clear in regards to what the
equivalent source is, and what it represents.
What it is: The equivalent source on the high
side is a voltage source in series with an
impedance.
5

What it represents: The equivalent source

represents the impedance between the high
side of the transformer and all of the current
sources in the remainder of the network.
Generally, the current sources in the
equivalent network are all of the generators.
20.2 Distributed generation
The above discussion assumes there are no
generators (current sources) connected along
the distribution feeder. This is usually the
case. However, if it is not the case, one
should be very careful to model those
generation units and capture their
contribution to the overall fault current.
During the past 5-10 years, there has been
growing interest in the topic of distributed
generation (DG). DG refers to the tendency
of generation to move from large,
centralized units connected at the
transmission or subtransmission level to
6

small, distributed units connected within the

distribution system.
Most peoples use of the term DG refers to
the following kinds of technologies, when
connected within the distribution system:
Reciprocating
(internal
combustion)
engines
Combustion turbines
Microturbines
Wind turbines
Fuel cells
Photovoltaics
One definition given for DG is utilization
of small (0 to 5 MW), modular power
generation
technologies
dispersed
throughout a utility s distribution system in
order to reduce T&D loading or load growth
and thereby defer the upgrade of T&D
facilities, reduce system losses, improve
power quality, and reliability.
7

Although DG has a lot of very attractive

effects, it causes major problems with
respect to distribution protection because of
the fundamental issue that its presence
causes fault currents to flow from multiple
directions. Relative to the radial, single
course configuration, design of protection is
considerably more complicated when DG is
present.
20.3 Three-phase analysis
We have seen in Notes7, that the positive,
negative, and zero sequence networks are
decoupled under the conditions that the
phase impedance matrix for all components
of the system have:
Equal
diagonal
elements
(phase
impedances must be equal), i.e.,
Z Z Z
(4)
Equal offdiagonal elements (offdiagonal
phase impedances must be equal), i.e.,
Z Z Z
(5)
aa

bb

cc

ab

ac

bc

We found in Notes7 that under the above

conditions, the 012 impedances are given by
Z Z Z Z Z Z 0
(6)
Z Z 2Z
(7)
Z Z Z Z
(8)
The resulting sequence networks are shown
in Fig. 4.
01

10

aa

02

20

12

21

ab

aa

ab

Fig. 4
We are very much aware that distribution
systems generally do not satisfy the
9

requirements in (4) and (5) above since

transposition is not used, that is, distribution
systems are not symmetric.
However, it is of some benefit to be aware
of the short circuit analysis procedures that
proceed from accepting the assumptions that
distribution systems are symmetric.
Since we always use the computer to obtain
an accurate answer, the reason for studying
use of the symmetrical components
approach for distribution systems is NOT
that it yields reasonably accurate answers
(although it sometimes does), but rather,
because it provide significant insight into
what happens in certain types of faulted
conditions.
The basic steps for using symmetrical
components in assessing faulted conditions
are as follows.
10

All quantities are assumed to be in per unit.

A. For positive, negative, & zero sequence:
1. Obtain the source impedance. Assume
that the voltage behind it is 1.0 per unit.
2.
Develop the sequence network for
the system under analysis.
3.
Obtain the Thevenin equivalent
looking into the network from the fault
point.
B. Connect the networks to capture the
influence of the particular fault type.
C. Compute the fault current from the
circuit resulting from step B.
D. From step C, you will also determine the
currents in all three of the networks
(positive, negative, and zero sequence
currents). This enables computation of the
phase currents Ia, Ib, and Ic from Iabc=AI012.
We discuss each one of these steps in what
follows.
11

20.4 Obtaining source impedance

There are several ways to do this [1], but the
easiest way is to obtain the positive,
negative, and zero-sequence Z-bus matrices
for the transmission network. The positive,
negative,
and
zero-sequence
source
impedances for the substation of interest is
the corresponding diagonal element in the
appropriate Z-bus. This is the impedance
seen from the bus looking back into the
system. These diagonal elements are also
called the driving point impedances.
If the HV bus of the distribution substation
is represented in the bus-impedance
matrices, then the diagonal elements must be
appropriately combined with the substation
transformer per-unit sequence impedances to
produce the source sequence impedances at
the low-voltage bus of the substation.

12

Normally, it is appropriate to assume that

negative sequence source impedance equals
positive sequence source impedance.
Depending on the transformer connection,
zero sequence source impedance may not be
necessary at all.
20.5 Developing sequence network
We address loads, lines, transformers, &
generators For each of these, we may derive
expressions for the 012 sequence
impedances via the following steps:
1. Express abc voltages as a function of abc
currents and abc impedances.
2. Substitute symmetric components for
abc voltages and currents (e.g., Vabc=AV012
and Iabc=AI012).
3. Manipulate to obtain the form
V012=Z012I012.
If you refer back to Notes7, you will see
that this is the procedure we followed to

13

obtain 012 sequence impedances for a Yconnected load.

However, we will not go through the
analytical details but will rather just state the
results.
20.5.1 Loads
Recall from Notes7 that for a Yconnected, balanced load grounded through
a neutral impedance Zn with impedance ZY
per phase, the 012 sequence circuits are as in
Fig. 5.

14

Fig. 5
where we see the sequence impedances are:
Z0=ZY+3Zn
(9)
Z1=ZY
(10)
Z2=ZY
(11)
If the neutral is solidly grounded, then Zn=0
and eq. (9) above becomes Z0=ZY.

15

If the neutral is ungrounded, then Zn=, and

eq. (9) above becomes Z0=, i.e., the 0sequence circuit is an open circuit, implying
no 0-sequence current flows in an
ungrounded Y connection.
For a delta-connected, balanced load, we
simply convert to an equivalent Y using
ZY=Z/3 and then apply relations for an
ungrounded Y connection, resulting in
Z0=
(12)
Z1= Z/3
(13)
Z2= Z/3
(14)
Example: A delta connected balanced load
with phase impedance Z is in parallel with
a solidly grounded Y-connected load with
phase impedance ZY. Draw the sequence
networks for the entire paralleled load.
It is possible to develop the sequence
networks using the 3-step approach given
above (and I have notes that do that).
16

However, this is very painful. Intuition,

which suggests that we should just obtain
the sequence networks of the parallel
combination as parallel combinations of the
individual sequence networks, is right.
Figure 6 shows the result.

Fig. 6

17

20.5.2 Lines
In Notes7, the work we did to answer the
question of What if the load (or line, or load
and line) is not symmetric? led to another
question, which was: So what are the
conditions for the off-diagonal elements of
Z012 to be 0?
We have already reviewed the answer to this
question in section 20.3 above, which was:
Equal
diagonal
elements
(phase
impedances must be equal), i.e.,
Z Z Z
(4)
Equal offdiagonal elements (offdiagonal
phase impedances must be equal), i.e.,
Z Z Z
(5)
In this case, the 012 sequence impedances
are:
Z Z Z Z Z Z 0
(6)
Z Z 2Z
(7)
Z Z Z Z
(8)
Equation (6) simply says that all offdiagonal elements on the 012 sequence
aa

bb

cc

ab

ac

bc

01

10

aa

02

20

12

21

ab

aa

ab

18

impedance matrix are zero. Equations (7)

and (8) provide the actual expressions that
we need for the 0, positive, and negative
sequence impedances.
The sequence networks are given in Fig. 7.

Fig. 7
It is interesting to compare eqs. (7) and (8)
for a symmetric line (or cable), and note that
19

they indicate that the zero-sequence quantity

is larger than the positive and negative
sequence quantities by 3Zab.
A typical overhead line has Zab(2/5)Zaa. In
this case, it is easy to show that Z 0=3Z1,
suggesting that finding zero sequence
impedance 3 times as large as positive
sequence impedance, is quite typical.
20.5.3 Transformers
There are five different types of transformer
connections to assess. These are:
1. Grounded Y to grounded Y.
2. Grounded Y to Y or Y to grounded Y.
3. -
4. Grounded Y to or to grounded Y
5. Y- or -Y.
As before, we can perform our 3-step
procedure given at the beginning of Section
20

20.5, where we express abc quantities,

substitute in symmetrical components, and
then obtain the decoupled equations. Here,
however, we must repeat this for both sides
of the transformer and then relate the two
sets of equations.
We will not perform this tedious work but
will instead simply observe some general
guidelines for drawing appropriate sequence
circuits. In forming these guidelines, we
assume:
Exciting current is negligible so shunt path
is infinite impedance and we only have the
series Z (winding resistance and leakage
reactance) in our transformer abc model.
Transformers in -Y or Y- configuration
are always connected so that positive
sequence voltages on the high side lead
positive sequence voltages on the low side
by 30 (per industry convention).
General guidelines for transformer 012
sequence circuits:
21

1. Positive
and
negative
sequence
impedances are equal, i.e.,
Z1=Z2=Zseries
where Zseries is the transformer winding
resistance and leakage reactance.
2. For connection types 4 and 5 in the
above list, the phase shift is included from
low side to high side as
+30 for positive sequence
-30 for negative sequence
3. For zero-sequence network,
a.We will get a complete open circuit
(I0=0) if there is an ungrounded Y on one
or both sides.
b. We will get isolation of primary from
secondary if there is a on one or both
sides. This means that connection
prevents pass-through of zero-sequence
currents. However, we may still get
zero-sequence current flowing if the
other side is grounded Y or .
c.We get no isolation if both sides are
grounded-Y.
22

d. Z0=Zseries+3Znp+3Zns where:
Znp: neutral impedance on primary
Zns: neutral impedance on secondary
A couple of concepts that are important here
in understanding the points under (3) above
are:
There must be a connection to ground on
the high-side for zero-sequence current to
flow from the transmission system into the
high side of the transformer.
If zero-sequence currents cannot flow
on primary (secondary) side of the
transformer, then because currents on the
secondary (primary) side of the
transformer can only arise through
induction of currents on the primary
(secondary) side of the transformer, zerosequence currents also cannot flow on
the secondary (primary) side of the
transformer.
23

So lets draw 012 sequence circuits for

various transformer connections.
1. Grounded Y to grounded Y.

Fig. 8

24

2. Grounded Y to Y or Y to grounded Y.

Fig. 9
Here, there is no place for zero-sequence
currents to flow on the Y side (since there is
no neutral and sum of phase currents, which
equals I0, must be 0). Therefore, there can be
no zero-sequence currents flowing on the
other side either. So I0=0 for this connection.
Y to grounded Y is the same.
25

3. -

Fig. 10
Here, zero sequence currents cannot enter or
leave either winding, although it is
possible for them to circulate within the
windings.

26

4. Grounded Y to or to grounded Y

Fig. 11: Grounded Y to

Fig. 12: to Grounded Y

27

In Figs. 11 and 12, we observe that zero

sequence currents can flow out of the
grounded Y side, which means they also
must be able to flow within but not out of
the .
We also observe that
-30 phase shift occurs from low side to
high side for positive sequence quantities
(which implies high side leads low side by
30 for positive sequence, in conformance
with industry convention)
30 phase shift occurs from low side to
high side for negative sequence quantities
(which implies high side quantities lag low
side by 30 for negative sequence, in
conformance with industry convention).

28

5.

Y- or -Y.

Fig. 13: Y to

29

Fig. 14: to Y
Observe that Figs. 13 and 14 are exactly like
Figs. 11 and 12 in the positive and negative
sequence circuits. The only difference in the
zero-sequence circuit, where we see that, in
Figs. 13 and 14, not only can zero-sequence
currents not pass through (which is the case
in Figs. 11 and 12) but they cannot flow at
all.
20.5.4 Generators

30

In distribution systems without any

distributed generation, so that all current
sources are upstream from the distribution
substation, the generation sequence circuits
are captured within the Z-bus, and so in such
cases, there is no real reason for distribution
analysis to deal with generation sequence
modeling.
But of course, if there is distributed
generation connected at the distribution
level, then we must deal with generation
sequence modeling.
We will assume for the sake of time that we
have no distributed generation modeled on
our distribution feeder, and so we will skip
discussion of quantities to use for generator
zero, positive, and negative sequence
impedances. (This subject, made interesting
by the fact that, unlike transformers and
lines, generators are rotating devices and
therefore positive sequence and negative
sequence quantities are different, is
31

addressed in most of the good power system

analysis textbooks).
One issue, however, is that because
generators produce balanced positive
sequence voltages, generators produce no
negative or zero sequence voltages.
Therefore, we only model a voltage source
in the positive sequence circuit.

Therefore, we model positive, negative, and

zero sequence circuits using our Z-bus
driving point impedances corresponding to
the transmission side of the distribution bus.
The positive and negative sequence
impedances are assumed equal, implying we
assume non-rotating devices (transformers
and lines) negative sequence impedance
dominates that of the rotating devices.

32

The sequence circuits we use in representing

the transmission network are given in Fig.
15.
When working in per-unit (as we have been
assuming throughout this discussion), the
positive sequence source voltage is typically
assumed to be Ean=1.0. Although it actually
may be something a little different than 1.0,
the influence on final short circuit currents
calculated is negligible.

33

Fig. 15
20.6 Obtaining Thevenin equivalent
The first thing to do is to draw all three
sequence networks of the system to be
analyzed, using information from the oneline diagram together with the component
sequence models provided in Figs. 5-15.
Then identify the fault location in each of
the three sequence networks.
Then, for each of the three sequence
networks,
determine
the
Thevenin
equivalent of the upstream portion of the
network as seen from the fault location.
Recall we
Obtain the Thevenin impedance by idling
all sources (short constant voltage sources
and open constant current sources) and

34

computing impedance see looking into the

network from the desired location, and
Obtain the Thevenin voltage by computing
the voltage at the open-circuited terminals
for the desired location.

Example:
Consider the substation of Fig. 16.

Fig. 16
The transformer is -Y grounded with rated
line-to-line ratio of 69 kV/12.47 kV and
power rating of 5 MVA. The transformer
35

positive sequence reactance is 0.065 pu on

its own base. The source reactance is 7.935
. Determine the positive, negative, and
zero-sequence Thevenin equivalents as seen
from Bus 2, on a 1 MVA base.
Solution:
The high-side impedance base is
ZB

69000 2
1E 6

4761

So the per-unit positive sequence source

impedance is
X S1

7.935
0.0017
4761

We will assume that the negative sequence

source reactance is the same.
The transformer reactance is 0.065 pu but it
is on a 5 MVA base. We need it on a 1 MVA
base. Therefore:
X T 1 0.065

1
0.0013
5

The sequence networks up to bus 2 are

drawn on the left-hand side of Fig. 17. Note
we have omitted the phase shifts for the -Y
transformer because we are only analyzing
36

currents on the low side. If we wanted

currents on the high side, we would need to
account for phase shifts in the positive and
negative sequence networks.

We draw them up to only bus 2 because that

is where we want to find the Thevenin
equivalents looking upstream (and so the
downstream network will not affect it).
The Thevenin equivalents, pictured on the
right of Fig. 17, are very easy to obtain,
which is typical for distribution networks.

37

Fig. 17

20.7 Connecting the networks

We will develop connections between the
sequence networks for each fault type and
apply them to our example.
20.7.1 Three-phase fault

38

A three phase fault has all three phases

connected to each other and to ground. It is a
symmetrical one; even with the fault, the
network is still symmetric.
To analyze it, we denote fault currents in
phases a, b, and c as Iaf/_0, Iaf/_-120,
Iaf/_120. Then the 012 currents are
computed as
1

I 012 A I abc
1 1
1
1 a
3
1 a 2

1 1
1
1 a
3
1 a 2

1 Ia
a 2 I b
a I c

1 I af 0
0

2
a I af 120 I af
0
a I af 120

This result shows what we could have stated

initially. Only positive sequence currents
flow for a 3-phase (symmetric) fault!
So we use only positive-sequence network
to analyze this fault. Fig. 18 shows the
network corresponding to a 3-phase fault.

39

Fig. 18
From Figure 18, we compute:
I 1fa

1.0
68.03 pu
0.0147

And of course,
I 0fa I 2fa 0

Now obtain the abc currents.

I fa

I 0fa

1 1

2
I fb A I 1 a
I fc
I
1 a

68.03

68.03 120
68.03120

1
fa
2
fa

1 0
a 68.03
a 2 0

So the a-phase fault current is 68.03 pu.

To get this in amperes, obtain the current
base on the low side of the transformer:
Ib

Sb 3
3VLLb

1E 6
3 (12.47 E 3)

46.3

Therefore the a-phase current for a threephase fault at bus 2 is 68.03(46.3)=3149.6A.

Conclusion: Analysis of a 3-phase fault
involves shorting the associated positive
sequence Thevenin network and computing
associated positive sequence current.
40

20.7.2 Single-phase fault

A single phase fault has one phase
connected to ground. It is also called a
single-line-to-ground (SLG) fault.
To analyze a SLG fault, denote fault currents
in phases a, b, and c as Iaf, 0, and 0. Then the
012 currents are computed as
1

I 012 A I abc
1 1
1
1 a
3
2
1 a

1 1
1
1 a
3
1 a 2

1 Ia
a 2 I b
a I c

I af
1 I af
1

2
a 0 I af
3
I af
a 0

We observe that the same current is flowing

in all three sequence networks! The circuit
connection for which this is true is a series
connection of the positive, negative, and
zero-sequence circuits.
We show the connected sequence circuits for
our example in Fig. 19.

41

Fig. 19
From Figure 19, we compute:
I 1fa I 2fa I 0fa

1.0
23.58 pu
0.0147 0.0147 0.013

Now obtain the abc currents.

I fa

I 0fa

1 1

2
I fb A I 1 a
I fc
I
1 a

70.74
0
0

1
fa
2
fa

1 23.58
a 23.58
a 2 23.58

So the a-phase fault current is 70.74 pu.

Recalling that the base current on the
secondary side is 46.3A, the a-phase current
for a SLG fault at bus 2 is
70.74(46.3)=3275.3A.
42

20.7.3 Line-to-line fault

A line-to-line fault has one phase connected
to another. Assuming phase b is connected to
phase c, we will have Iaf=0, Icf=- Ibf.
To analyze a SLG fault, denote fault currents
in phases a, b, and c as Iaf, 0, and 0. Then the
012 currents are computed as
1

I 012 A I abc
1 1
1
1 a
3
1 a 2

1 1
1
1 a
3
1 a 2

1
a 2
a

1
a 2

Ia
I
b

a I c

0
0
1

I bf
3 I bf 90
3
3 I bf 90
I bf

We observe that the zero sequence circuit is

dead, but positive and negative sequence
currents are equal in magnitude but opposite
in dissection! The circuit connection for
which this is true is one where the positive
sequence terminals are directly connected to
the negative sequence terminals.

43

We show the connected sequence circuits for

our example in Fig. 20.

Fig. 19
From Figure 19, we compute:
I 1fa I 2fa

1.0
34.01 pu
0.0147 0.0147

Now obtain the abc currents.

I fa
I 0fa 1 1

1
2
I fb A I fa 1 a
I fc
I 2fa 1 a

3 ( 34.01)90
3 ( 34.01) 90

1
a
a 2

0
34.01

34.01
0

58.91
58.91

So the a-phase fault current is 58.91 pu.

Recalling that the base current on the
secondary side is 46.3A, the a-phase current
for a SLG fault at bus 2 is
58.91(46.3)=2727.5A.
44

20.7.4 Line-to-line to ground fault

References
[1] McGraw-Edison Power Systems,
Distribution system protection manual.
[2] T. Short, Electric power distribution
handbook, CRC press, 2004.
[3] T. Gonen, Electric Power Distribution
System Engineering, McGraw-Hill, 1986.

45