ECE 476

Power System Analysis

Lecture 21: Unbalanced Faults, Protection

Prof. Tom Overbye

Dept. of Electrical and Computer Engineering
University of Illinois at Urbana-Champaign
overbye@illinois.edu
Announcements

• Please read Chapters 9, 10 and 11

• HW 8 quiz today
• HW 9 is 8.14, 9.1, 9.2 (bus 3), 9.15, 9.64; this should
be turned in on Nov 10 (hence no quiz)
• Chapter 6 Design Project 1 is assigned. It will count as
three regular home works and is due on Dec 3.
– For tower configurations assume a symmetric conductor spacing, with the distance
in feet given by the following formula: (Last two digits of your EIN+150)/10.
Example student A has an UIN of xxx65. Then his/her spacing is (65+150)/10 =
21.50 ft.

• Exam 2 is during class on Tuesday November 15

• Final exam is on Monday December 12, 1:30-4:30pm
1
Unbalanced Fault Analysis

• The first step in the analysis of unbalanced faults is

to assemble the three sequence networks. For
example, for the earlier single generator, single
motor example let’s develop the sequence networks

2
Sequence Diagrams for Example

Positive Sequence Network

Negative Sequence Network

3
Sequence Diagrams for Example

Zero Sequence Network

4
Create Thevenin Equivalents

• To do further analysis we first need to calculate the

thevenin equivalents as seen from the fault location.
In this example the fault is at the terminal of the right
machine so the thevenin equivalents are:

Zth  j 0.2 in parallel with j0.455

Zth  j 0.21 in parallel with j0.475
5
Single Line-to-Ground (SLG) Faults

• Unbalanced faults unbalance the network, but only

at the fault location. This causes a coupling of the
sequence networks. How the sequence networks
are coupled depends upon the fault type. We’ll
derive these relationships for several common
faults.
• With a SLG fault only one phase has non-zero fault
current -- we’ll assume it is phase A.

6
SLG Faults, cont’d

 I af   ? 
 f  
 Ib   0 
 f  0 
 I c   

Then since
 I 0f  1 1 1  ? 
  1
 2    1 f
 I f   1    0  I f  I f  I f  I a
0
  3   3
 1  2
  0 
 I f   
7
SLG Faults, cont’d

Vaf  Z f I af
Vaf  1 1 1 f V 0

 f     

Vb   1  2
  f 
 V
 f  1   2    
Vc    V f 
This means Vaf  V f0  V f  V f
The only way these two constraints can be satisified
is by coupling the sequence networks in series

8
SLG Faults, cont’d

With the
sequence
networks in
series we can
solve for the
fault currents
(assume Zf=0)
1.050
I f    j1.964  I f  I 0f
j (0.1389  0.1456  0.25  3Z f )
I  A I s  I af   j 5.8 (of course, Ibf  I cf  0)
9
Problem 9.59 Example
Line-to-Line (LL) Faults

• The second most common fault is line-to-line,

which occurs when two of the conductors come in
contact with each other. With out loss of generality
we'll assume phases b and c.

Current Relationships: I af  0, Ibf   I cf , I 0f  0

Voltage Relationships: Vbg  Vcg

11
LL Faults, cont'd

Using the current relationships we get

 I 0f  1 1 1  0 
 
 1 2 f 
 I f   1     I b 
  3
1  2     I f 

 I f    b 
I 0f  0

I f
1 f

 Ib    2
3
 I f 
1 f 2
 Ib   
3

Hence I f   I f

12
LL Faults, con'td

Using the voltage relationships we get

V f0  1 1 
1  Vag
f 
  1   

V f   1   2  Vbgf  Recall during
  3
1    this fault the

V f   2   Vcgf 
  b and c phase
Hence voltages
are identical
 1 f
V f  Vag 
3
 
   2 Vbgf 

 1 f
V f  Vag 
3
 
 2   Vbgf   V f  V f

13
LL Faults, cont'd

To satisfy I f   I f & V f  V f
the positive and negative sequence networks must
be connected in parallel

There is no
zero sequence
current because
there is no
path to ground

14
LL Faults, cont'd

Solving the network for the currents we get

 1.050
If   3.691  90
j 0.1389  j 0.1456
 I af  1 1 1  0   0 
 f   
 Ib   1 
2
  3.691  90   6.39 
   
 f  1   2   3.69190   6.39 
 I c   

15
LL Faults, cont'd

Solving the network for the voltages we get

V f  1.050  j 0.1389  3.691  90  0.5370

Vf   j 0.1452  3.69190  0.5370
Vaf  1 1 1   0   1.074 
 f   
Vb   1 
2
  0.537    0.537 
   
 f  1   2  0.537   0.537 
Vc   

16
Double Line-to-Ground Faults

• With a double line-to-ground (DLG) fault two line

conductors come in contact both with each other
and ground. We'll assume these are phases b and c.

Iaf
0 f
Vbg  Vcg
f
 Z f ( Ibf
 Icf
)
17
DLG Faults, cont'd

From the current relationships we get

 I af  1 1 
1  If
0

 f    
 Ib   1  2
  I f 
 f  1  2 
 I c     I f 
 
Since I af  0  I 0f  I f  I f  0
Note, because of the path to ground the zero
sequence current is no longer zero.

18
DLG Faults, cont'd

From the voltage relationships we get

V f0  1 1 
1  agV f 
  1    
V f   1   2  Vbgf  
  3
1  2    f 

V f    Vbg 
 
Since f
Vbg  Vcg
f
 Vf  Vf
Then Vbgf  V f0  ( 2   )V f
But since 1     2  0   2    1
Vbgf  V f0  V f
19
DLG Faults, cont'd

Vbgf  V f0  V f
 Z f ( I bf  I cf )
Also, since
I bf  I 0f   2 I f   I f
I cf  I 0f   I f   2 I f
Adding these together (with    2  -1)
   
f
Vbg  0
Z f (2 I f  If  If ) with 0
If  I f  If
V f0  V f  3Z f I 0f
20
DLG Faults, cont'd

• The three sequence networks are joined as

follows

Assuming Zf=0, then


 V 1.050
If   

Z  Z ( Z  3Z f ) j 0.1389  j 0.092
0

 4.547  0
21
DLG Faults, cont'd

V f  1.05  4.547  90  j 0.1389  0.4184

I f  0.4184 / j 0.1456  j 2.874
I 0f   I f  I f  j 4.547  j 2.874  j1.673
Converting to phase: I bf  1.04  j 6.82
I cf  1.04  j 6.82
22
Unbalanced Fault Summary

• SLG: Sequence networks are connected in series,

parallel to three times the fault impedance
• LL: Positive and negative sequence networks are
connected in parallel; zero sequence network is not
included since there is no path to ground
• DLG: Positive, negative and zero sequence
networks are connected in parallel, with the zero
sequence network including three times the fault
impedance

23
Power System Protection

• Main idea is to remove faults as quickly as possible

while leaving as much of the system intact as
possible
• Fault sequence of events
– Fault occurs somewhere on the system, changing the
system currents and voltages
– Current transformers (CTs) and potential transformers
(PTs) sensors detect the change in currents/voltages
– Relays use sensor input to determine whether a fault has
occurred
– If fault occurs relays open circuit breakers to isolate fault

24
Power System Protection

• Protection systems must be designed with both

primary protection and backup protection in case
primary protection devices fail
• In designing power system protection systems there
are two main types of systems that need to be
considered:
1. Radial: there is a single source of power, so power
always flows in a single direction; this is the easiest from
a protection point of view
2. Network: power can flow in either direction: protection
is much more involved

25
Radial Power System Protection

• Radial systems are primarily used in the lower

voltage distribution systems. Protection actions
usually result in loss of customer load, but the
outages are usually quite local.
The figure shows
potential protection
schemes for a
radial system. The
bottom scheme is
preferred since it
results in less lost load.
26
Radial Power System Protection

• In radial power systems the amount of fault current is

limited by the fault distance from the power source:
faults further down the feeder have less fault current
since the current is limited by feeder impedance
• Radial power system protection systems usually use
inverse-time overcurrent relays.
• Coordination of relay current settings is needed to
open the correct breakers

27
Inverse Time Overcurrent Relays

• Inverse time overcurrent relays respond instan-

taneously to a current above their maximum setting
• They respond slower to currents below this value
but above the pickup current value

28
Inverse Time Relays, cont'd

• The inverse time characteristic provides backup

protection since relays further upstream (closer to
power source) should eventually trip if relays
closer to the fault fail
• Challenge is to make sure the minimum pickup
current is set low enough to pick up all likely
faults, but high enough not to trip on load current
• When outaged feeders are returned to service there
can be a large in-rush current as all the motors try
to simultaneously start; this in-rush current may re-
trip the feeder
29
Inverse Time Overcurrent Relays
Current and time
settings had been
adjusted using dials
on the relay
Relays have
traditionally been
electromechanical
devices, but are
gradually being
replaced by
digital relays
30