Fault detection in transmission networks of power systems

S. Saha
a,
, M. Aldeen
a
, C.P. Tan
a,b
a
Department of Electrical and Electronic Engineering, The University of Melbourne, Parkville, Victoria 3010, Australia
b
School of Engineering, Monash University, Jalan Lagoon Selatan, 46150 Bandar Sunway, Malaysia
a r t i c l e i n f o
Article history:
Received 4 November 2009
Received in revised form 8 December 2010
Accepted 9 December 2010
Available online 18 February 2011
Keywords:
SMIB
Synchronous generator
Fault detection
Modelling
Sliding mode observer
Transmission lines
a b s t r a c t
An online fault detection scheme for a sample power system is introduced in this paper. The detection
approach is based on the use of a variable structure system called sliding mode observer, where infor-
mation contained in the output measurements is utilized to detect the onset of faults in the transmission
network of the sample power system in real time and online. The power system comprises a generating
unit connected to an innite bus through double line transmission network. The common case of a fault
occurring on a transmission line is illustrated through simulation studies.
2011 Elsevier Ltd. All rights reserved.
1. Introduction
Power systems are prone to frequent faults, which may occur in
any of its components, such as generating units, transformers,
transmission network and/or loads. It is well known that faults
can cause signicant disruption of supply, destabilise the entire
system and may also cause injuries to personnel. Detection of
faults is therefore of a paramount importance from economic and
operational viewpoints. In addition faults should be detected as
quickly as possible, in real time if possible, so that an appropriate
remedial action can be promptly taken before major disruptions
to the power supply can occur.
So far such faults are detected by fault locators, which form an
integral part of protection systems. However, the accuracy of
fault locators is known to be below desirable levels. It is also
known that fault locators may not able to detect faults in real time.
In this paper we introduce a software based scheme where faults in
the transmission network are able to be detected in real time, using
commonly available measurements of speed, load angle, terminal
voltage, power, etc. Only symmetrical faults are considered here,
as asymmetrical faults require dynamical incorporation of sym-
metrical components into the overall dynamical systems, which
is beyond the scope of this paper. However research in this direc-
tion by the authors is ongoing and newresults will soon be submit-
ted for possible publication.
The novelty of the proposed scheme is two folds; development
of a fault dependent model for a sample power system and design
of a real time fault detectors, referred to in the control literature as
fault detection lter. The result of this study are among a very
few model based results that have been reported in the open liter-
ature (see for example [1,2]) and may therefore offer tangible ben-
ets to the protection industry if incorporated in the design of fault
locators. It can be noted that this paper does not deal with protec-
tion systems per say but with how to locate faults in real time
using control and estimation theory. Protection systems experts
are therefore invited to make use of this approach.
Other reported contributions are reported in [1618]. In [16],
pre-fault and during fault phasor currents and voltages are mea-
sured and used in a two-bus Thevenin equivalent network model
of the transmission system to locate a fault after its occurrence.
In this method the dynamics of the system are ignored and each
generator is represented by a voltage behind transient reactance.
In [17], adaptive extended Kalman lter and probabilistic neural
networks are combined to estimate different harmonic compo-
nents in fault current signals. The harmonic components are then
fed into a forward neural network for training and identication
of high impedance faults in power distribution feeders off line. In
[18], discrete wavelet transform and neural networks are used to
parameterise and characterise fault signals.
Unlike the approach presented in this paper, the above ap-
proaches are not suitable for real time fault detection as they are
based on acquiring fault date, which can then be used in knowl-
edge based techniques for analysis and characterisation.
0142-0615/$ - see front matter 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijepes.2010.12.026

Corresponding author. Tel.: +61 383440374.

E-mail address: s.saha@pgrad.unimelb.edu.au (S. Saha).
Electrical Power and Energy Systems 33 (2011) 887900
Contents lists available at ScienceDirect
Electrical Power and Energy Systems
j our nal homepage: www. el sevi er . com/ l ocat e/ i j epes
This paper is structured as follows: We start with the develop-
ment of a comprehensive model of a power system suitable for
fault studies and test its validity through simulation studies. Then
we introduce a software based fault detection algorithm (lter)
capable of detecting faults online and real time. The fault lter is
then applied to the sample power system to detect faults in the
transmission network. We nally present comprehensive simula-
tion studies carried out on the power system under various scenar-
ios and discuss the results in full.
2. Generating unit model
In the following we provide a detailed linearised model of a
generating unit comprising an 8th order synchronous generator,
a 3rd order model of an IEEE Type 1S exciter [3], and a 3rd order
model of an IEEE steam-turbine speed governor [4]. These models
are based on those presented in with some changes.
2.1. Synchronous generator
Detailed models of a synchronous generator should include
electrical (ux-voltage equations) and mechanical (rotor shaft mo-
tion) parts [57]. We start with modelling the ux-voltage dynam-
ical relationships along the d-axis. These relationships may be
modelled by the set of linear equations listed below.
_
w
d
x
b
w
q
1
_
d

x
b
w

q
_
d x
b
v
d
x
b
ri
d
1
_
w
F
x
b
v
F
r
F
i
F
x
b
2
_
w
D
x
b
v
D
x
b
r
D
i
D
3
In the q-axis the voltage-ux dynamics are described by the follow-
ing linear equations:
_
w
q
x
b
w
d
1
_
d

x
b
w

d
_
d x
b
v
q
x
b
ri
q
4
_
w
G
x
b
v
G
x
b
r
G
i
G
5
_
w
Q
x
b
v
Q
x
b
r
Q
i
Q
6
The mechanical part of the machine is due to the motion of the rotor
shaft and is expressed as follows:

d
x
b
2H
k
d
_
d T
m
T
e

x
b
2H
7
T
e
w
d
i

q
w

d
i
q
w
q
i

d
w

q
i
d
8
2.2. Excitation system
The synchronous machine is equipped with the following IEEE
excitation system.
The state equations for the excitation system shown in Fig. 1 are
derived as follows:
_
V
R

1
s
R
V
R

K
R
s
R
jV
N
j 9
_
V
F

1
s
F
V
F

K
F
s
F
_
V
A
10
_
V
A

1
s
A
V
A

K
A
s
A
V
REF
V
R
V
F
11
2.3. IEEE governor model
A model of a steam turbine speed governing system used in this
study is shown in Fig. 2 and its model is derived below:
_
P
ST

1
T
ST
P
ST

1
T
ST
P
GV
12
_
P
GV

1
T
SM
P
GV

1
T
SM
P
SR
13
_
P
SR

1
T
SR
P
SR

1
T
SR
P
r

_
d
Rx
B
_ _
14
In order to interface the governor with the synchronous generator,
the governor output, which is the mechanical power, P
m
, needs to
Nomenclature
w ux linkage
.
0
superscript refers to nominal value
.
q
subscript refers to the quadrature axis
.
G,Q
subscript refers to damper windings on the quadrature
axis
.
d
subscript refers to the direct axis
.
D
subscript refers to damper winding on the direct axis
.
F
subscript refers to eld winding
.
ex
subscript refers to exciter
.
N
subscript refers to network
.
G
subscript refers to governor
x state vector
y output vector
f fault parameters
A, B, C, D, E, G, W constant parameter matrices
d rotor shaft angle
x rotor shaft speed
x
b
speed base quantity
v, V machine and network voltage
i, I machine and network current
jj magnitude of a vector
E
FD
eld winding excitation voltage
H synchronous machine inertia constant
k
d
rotor damping coefcient
T
m
, P
m
mechanical torque and power
REF
V
FD
E
A
V
1
R
R
K
s +
1
A
A
K
s +
1
F
F
sK
s +
+
-
V
V
2
V
F
|V
N
|
-
R
+
Fig. 1. IEEE type 1 excitation system.
Fig. 2. IEEE steam turbine speed governing system.
888 S. Saha et al. / Electrical Power and Energy Systems 33 (2011) 887900
be expressed in terms of mechanical torque, T
m
, to act as an input to
the generator. This is accomplished as follows. The relationship be-
tween the mechanical torque and the power is given as:
P
m
T
m
x T
m
1
_
d
Linearization of this equation and re-arranging will give
T
m

P
m
T

m
_
d
1
_
d

15
3. Transmission network model
A Single Machine Innite Bus (SMIB) power system is shown in
Fig. 3, where possible three phase line-to-ground faults, denoted
by f
1,2
, on the two lines are also shown. The location of the fault
is at a distance g
1,2
from the generator terminal. In this study only
symmetrical faults are considered. Asymmetrical faults require
dynamical incorporation of symmetrical components into the
overall dynamical systems, which is beyond the scope of this pa-
per. New results in this respect will soon be submitted for possible
publication.
From Fig. 3, the individual line currents are expressed as
I
N1

f
1
g
1
Y
1
V
N
1 f
1
Y
1
V
N
V
B

I
N2

f
2
g
2
Y
2
V
N
1 f
2
Y
2
V
N
V
B

I
N
I
N1
I
N2
16
From (16) it is easy to derive a generic expression for a SMIB system
that has as many as lines as required, say k lines, as follows:
V
N

k
i1
f
i
g
i
Y
i
f1 f
i
gY
i
_ _
_ _
1
I
N

k
i1
1 f
i
Y
i
V
B
_ _
17
For a two-line conguration Eq. (17) is written as
V
N

f
1
g
1
Y
1
f1 f
1
gY
1

f
2
g
2
Y
2
f1 f
2
gY
2
_ _
1
fI
N
1 f
1
Y
1
V
B
1 f
2
Y
2
V
B
g
Or in a more compact form as
V
N

1
g
1
1
_ _
Y
1
1
g
2
1
_ _
Y
2
_ _
f
1
f
2
_ _
Y
1
Y
2
_ _
1
I
N
Y
1
Y
2
V
B
Y
1
Y
2
V
B
f
1
f
2
_ _ _ _
18
Therefore Eq. (18) can simulate any possible scenario of faults
occurring on either line and at any distance from the generator. It
may also be expressed in a vector form as shown below:
V
N
l
1
l
2

f
1
f
2
_ _
Y
12
_ _
1
I
N
Y
1
Y
2
V
B
Y
1
Y
2
V
B
f
1
f
2
_ _ _ _
19
where
l
1

1
g
1
1
_ _
Y
1
; l
2

1
g
2
1
_ _
Y
2
; Y
12
Y
1
Y
2
Or
V
N
lf Y
12

1
fI
N
Y
12
V
B
Y
f
V
B
f g 20
where
l l
1
l
2
; f
f
1
f
2
_ _
; Y
12
Y
1
Y
2
; Y
f
Y
1
Y
2

It is clear that Eq. (20) represents any number of faults that may oc-
cur in any transmission network. It however can be simplied if we
use the fact that the two lines, 1 and 2, are the same. Then we can
write
Y Y
1
Y
2
R jX
1
21
Using Eq. (21) in Eq. (20), we obtain the following fault cases.
No fault case: For no fault condition, i.e. f
1
= f
2
= 0, Eq. (20) is re-
duced to the nominal case expressed as
V
N

1
2
Y
1
I
N
V
B
22
Fault on line 1: In case of fault occurring on line 1, for example, i.e.
f
1
= 1, but line 2 is fault-free then Eq. (20) reduces to
V
N
1
1
g
1
_ _
Y
_ _
1
fI
N
YV
B
g 23
Fault on line 2: Similarly, in case of fault occurring on line 2, for
example, i.e. f
2
= 1, but line 1 is fault-free then Eq. (20) reduces to
V
N
1
1
g
2
_ _
Y
_ _
1
fI
N
YV
B
g 24
Faults on line 1 and line 2: In case of fault occurring on both line 1
and line 2, for example, i.e. f
1
= f
2
= 1 then Eq. (20) reduces to
Fig. 3. SMIB power system. Fig. 4. Phasor relationship between V
B
and V
N
.
S. Saha et al. / Electrical Power and Energy Systems 33 (2011) 887900 889
0 5 10 15 20 25 30 35 40
Load Angle Response
Speed Response
0
-0.5
0.5
Terminal Voltage
0 5 10 15 20 25 30 35 40
-1
-0.5
0
0.5
Accelerating Torque Response
0 5 10 15 20 25 30 35 40
-1
-0.5
0
0.5
Terminal Power Response
0 5 10 15 20 25 30 35 40
-4
-2
0
2
0 5 10 15 20 25 30 35 40
Time (Seconds)
-2
0
2
4
Fig. 5. Output responses to a 5% change in V
REF
.
0 10 20 30 40 50 60
0 2 4 6 8 10 12 14 16 18
L
o
a
d

A
n
g
l
e
L
o
a
d

A
n
g
l
e
15 20 25 30 35 40 45 50 55 60
L
o
a
d

A
n
g
l
e
Time (Seconds)
-40
-20
0
20
-1
-0.5
0
0.5
-40
-20
0
Fig. 6. Load angle response.
890 S. Saha et al. / Electrical Power and Energy Systems 33 (2011) 887900
V
N

1
g
1

1
g
2
_ _
Y
_ _
1
I
N
:
Finally, we may re-write Eq. (20) in a state space form by dening
b
i
1
1
g
i
_ _
1
and manipulating to obtain.
V
N

1
2
b
1

1
2
_ _
b
2

1
2
_ _ _ f
1
f
2
_ _ _ _
Y
1
I
N
1 b
1
1 b
2
1
f
1
f
2
_ _ _ _
V
B
25
Eq. (25) represents a generic fault-dependent state space expression
that can be easily incorporated into the overall SMIB dynamical
model, as outlined below.
As shown in Fig. 1, the input to the exciter is the magnitude of
the terminal voltage. In order to incorporate this into the overall
generating unit model it is necessary to represent the machine ter-
minal voltage in terms of the generating unit state variables. This is
accomplished next.
A phasor diagram representing this relationship is shown in
Fig. 4, where the load angle d is one of the state variables, dened
in Section 2 above and the magnitude of V
B
is commonly assumed
constant, 1 pu. From the gure, the components of V
B
on the d- and
q-axes, are expressed as
v
Bq
jV
B
j cos d; v
Bd
jV
B
j sind
Thus taking the d-axis as the reference axis, we write
V
B
v
Bd
jv
Bq
26
Similarly, we write
0 10 20 30 40 50 60
S
p
e
e
d
0 2 4 6 8 10 12 14 16 18
S
p
e
e
d
15 20 25 30 35 40 45 50 55 60
S
p
e
e
d
Time (seconds)
-4
-2
0
2
-1
-0.5
0
0.5
-4
-2
0
2
Fig. 7. Speed response.
0 10 20 30 40 50 60
15 20 25 30 35 40 45 50 55 60
Time
V
n
-1
-0.5
0
0.5
V
n
-1
-0.5
0
0.5
V
n
-1
-0.5
0
0.5
0 2 4 6 8 10 12 14 16 18
Fig. 8. Terminal voltage response.
S. Saha et al. / Electrical Power and Energy Systems 33 (2011) 887900 891
V
N
v
d
jv
q
27
Equating Eq. (27) with Eq. (25), substituting for I
N
= i
d
+ ji
q
,
V
B
= v
Bd
+ jv
Bq
and using Y = R + jX yields
v
d
jv
q

1
2
b
1

1
2
_ _
b
2

1
2
_ _ _ f
1
f
2
_ _ _ _
R jXi
d
ji
q

1 b
1
1 b
2
1
f
1
f
2
_ _ _ _
v
Bd
jv
Bq

_
_

_
28
From Eq. (28), the real and imaginary parts of the terminal voltage
are derived as
v
d

1
2
b
1

1
2
_ _
b
2

1
2
_ _ _ f
1
f
2
_ _ _ _
Ri
d
Xi
q

1 b
1
1 b
2
1
f
1
f
2
_ _ _ _
jV
B
j sind
_

_
_

_
29
v
q

1
2
b
1

1
2
_ _
b
2

1
2
_ _ _ f
1
f
2
_ _ _ _
Ri
q
Xi
d

1 b
1
1 b
2
1
f
1
f
2
_ _ _ _
jV
B
j cos d
_

_
_

_
30
Linearising Eqs. (29) and (30) gives, respectively
0 5 10 15 20 25 30 35 40

, load angle
d / dt, load speed
T
a
, accelerating torque
P
N
, machine terminal power
-0.5
0
0.5
0 5 10 15 20 25 30 35 40
V
N
, terminal voltage of generator
-5
0
5
-5
0
5
0 5 10 15 20 25 30 35 40
0 5 10 15 20 25 30 35 40
0 5 10 15 20 25 30 35 40
-10
0
10
-10
0
10
Fig. 9. The response of the outputs when V
REF
is increased.
892 S. Saha et al. / Electrical Power and Energy Systems 33 (2011) 887900
v
d

jV
B
j cos d

d
1
2
Ri
d

1
2
Xi
q

b
1

1
2
_ _
Ri

d
b
1

1
2
_ _
Xi

q
b
1
1jV
B
j sind

_ _
b
2

1
2
_ _
Ri

d
b
2

1
2
_ _
Xi

q
b
2
1jV
B
j sind

_ _ _ _
f
1
f
2
_ _
_
_
_
_
_
_
_
_
31
v
q

jV
B
j sind

d
1
2
Ri
q

1
2
Xi
d

b
1

1
2
_ _
Ri

q
b
1

1
2
_ _
Xi

d
b
1
1jV
B
j cos d

_ _
b
2

1
2
_ _
Ri

q
b
2

1
2
_ _
Xi

d
b
2
1jV
B
j cos d

_ _ _ _
f
1
f
2
_ _
_
_
_
_
_
_
_
_
32
Finally the magnitude of the machine terminal voltage is now ob-
tained from
jV
N
j

v
2
d
v
2
q
_
33
A linearised form of the machine terminal voltage is derived as
jDV
N
j
v

d
Dv
d
v

q
Dv
q

v
2
d
v
2
q
_ 34
dening a
c
1=

v
2
d
v
2
q
_
and dropping D results in
z
1
and its estimate (dotted)
z
2
and its estimate (dotted)
0 5 10 15 20 25 30 35 40
z
3
and its estimate (dotted)
0 5 10 15 20 25 30 35 40
z
4
and its estimate (dotted)
0 5 10 15 20 25 30 35 40
z
5
and its estimate (dotted)
0 5 10 15 20 25 30 35 40
0 5 10 15 20 25 30 35 40
-4
-2
0
2
-20
-10
0
10
-10
-5
0
5
-1
0
1
2
-0.5
0
0.5
Fig. 10. The response of the outputs of the augmented system and the corresponding observer outputs (dotted) when V
REF
is increased.
S. Saha et al. / Electrical Power and Energy Systems 33 (2011) 887900 893
jVNj
ac v

d
jVBj cos d

d
1
2
Ri
d

1
2
Xiq
b
1

1
2
_ _
Ri

d
b
1

1
2
_ _
Xi

q
b
1
1jVBj sind

_ _
b
2

1
2
_ _
Ri

d
b
2

1
2
_ _
Xi

q
b
2
1jVBj sind

_ _ _ _
f1
f2
_ _
_
_
_
_
_
_
_
_
ac v

q
jVBj sind

d
1
2
Riq
1
2
Xi
d

b
1

1
2
_ _
Ri

q
b
1

1
2
_ _
Xi

d
b
1
1jVBj cos d

_ _
b
2

1
2
_ _
Ri

q
b
2

1
2
_ _
Xi

d
b
2
1jVBj cos d

_ _ _ _
f1
f2
_ _
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
35
In order to express the terminal voltage in terms of the state vari-
ables of the generating unit, we rst need to express the currents
in terms of the state variables as outlined below:
I
x
L
1
x
w
and L
L
d
L
ad
L
ad
0 0 0
L
ad
L
F
L
ad
0 0 0
L
ad
L
ad
L
D
0 0 0
0 0 0 L
q
L
aq
L
aq
0 0 0 L
aq
L
G
L
aq
0 0 0 L
aq
L
aq
L
Q
_

_
_

_
36
where
I
x
i
d
i
F
i
D
i
q
i
G
i
Q

T
Hence we can write
i
d
e
1
L
1
x
w
i
q
e
4
L
1
x
w
37
where
e
1
1 0 0 0 0 0; e
4
0 0 0 1 0 0
As a result Eq. (35) can be expressed as
jVNj
ac v

d
jVBj cos d

d
1
2
Re1L
1
xw
1
2
Xe4L
1
xw
b
1

1
2
_ _
Ri

d
b
1

1
2
_ _
Xi

q
b
1
1jVBj sind

_ _
b
2

1
2
_ _
Ri

d
b
2

1
2
_ _
Xi

q
b
2
1jVBj sind

_ _ _ _
f1
f2
_ _
_
_
_
_
_
_
_
_
_
_
ac v

q
jVBj sind

d
1
2
Re4L
1
xw
1
2
Xe1L
1
xw
b
1

1
2
_ _
Ri

q
b
1

1
2
_ _
Xi

d
b
1
1jVBj cos d

_ _
b
2

1
2
_ _
Ri

q
b
2

1
2
_ _
Xi

d
b
2
1jVBj cos d

_ _ _ _
f1
f2
_ _
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
38
Eq. (38) can be expressed in a matrix-vector compact form as
jV
N
j G
w
G
d
x
T
w
x
T
d
_
T
G
f
f 39
where
G
w

a
c
2
v

d
R v

q
X
_ _
e
1
L
1
v

q
R v

d
X
_ _
e
4
L
1
_ _
40
G
d
a
c
v

d
V
B
cos d

a
c
v

q
V
B
sind

_ _
0
_ _
41
G
f

a
c
v
0
d
b
1

1
2
_ _
Ri

d
b
1

1
2
_ _
Xi

q
b
1
1jV
B
j sind

_ _
a
c
v
0
q
b
1

1
2
_ _
Ri

q
b
1

1
2
_ _
Xi

d
b
1
1jV
B
j cos d

_ _
a
c
v
0
d
b
2

1
2
_ _
Ri

d
b
2

1
2
_ _
Xi

q
b
2
1jV
B
j sind

_ _
a
c
v
0
q
b
2

1
2
_ _
Ri

q
b
2

1
2
_ _
Xi

d
b
2
1jV
B
j cos d

_ _
_

_
_

_
T
42
4. Derivation of state space model
4.1. State equations
A linearised model of the generating unit can now be expressed,
in state space form, by combining the synchronous machine, exci-
ter, and governor models. As a result the following state equation is
obtained
_ x Ax Bu E
x
f 43
where the state vector is dened as
x x
T
w
x
T
d
x
T
ex
x
T
gov
_
T
; x
w
w
d
w
F
w
D
w
q
w
G
w
Q

T
;
x
d
d
_
d
T
;
x
ex
V
R
V
F
V
A

T
; x
gov
P
ST
P
GV
P
SR

T
and the input vector u is dened as.
u V
REF
P
REF

T
V
REF
is the reference voltage and P
REF
is the reference power. The
matrix parameters A, B, E
x
are derived in Appendix A.
4.2. Output equations
For the purposes of simulation, state estimation and controller
design, it is commonto use the most easily available measurements,
such as load angle, speed, terminal voltage, accelerating torque and
terminal power. We therefore dene our output equation as
y d
_
d V
N
T
a
P
N

T
Cx E
y
f 44
The parameters C, E
y
are derived in Appendix B.
0 10 20 30 40
The fault
0
0.2
0.4
0.6
0.8
1
0 10 20 30 40
The fault reconstruction
0
0.2
0.4
0.6
0.8
1
Fig. 11. The fault (left) and its reconstruction (right) when V
REF
is increased.
894 S. Saha et al. / Electrical Power and Energy Systems 33 (2011) 887900
5. Fault studies
In this section, we simulate cases fault conditions for the linear
power system described above. A number of scenarios are consid-
ered where fault occur on one line only and on both lines at differ-
ent locations.
5.1. Case study I: fault on a single line
This study involves the following sequence of events. First we
run the system unperturbed (i.e. no changes in the reference inputs
and no fault). Second, at time t = 2 s, we increase the reference volt-
age V
REF
by 5% and run the simulation for another 25 s. Third, we
inject three-phase line-to-ground fault on line 2 of the transmis-
sion system at time 25 s.
The output responses are shown in Fig. 5. It is clear from the
individual gures that the load angle slips back as a result of the
increase in the terminal voltage. This is due to the fact that as
the voltage increases, the terminal power increases and as the in-
put power remains constant, the increase in the terminal power is
met by release of the kinetic energy stored in the rotor shaft, hence
the slipping back of the rotor angle. The slipping back of the load
angle is accompanied by a decrease in the speed, but as the unit
is connected to an innite bus, the speed recovers to that of the
innite bus and therefore no change in the speed occurs at steady
state.
0 5 10 15 20 25 30 35 40
0 5 10 15 20 25 30 35 40
0 5 10 15 20 25 30 35 40
, load angle
-4
-2
0
2
d / dt, load speed
-4
-2
0
2
V
N
, terminal voltage of generator
T
a
, accelerating torque
0 5 10 15 20 25 30 35 40
-5
0
5
10
0 5 10 15 20 25 30 35 40
-10
-5
0
5
P
N
, machine terminal power
-0.5
0
0.5
Fig. 12. The response of the outputs when P
REF
is increased.
S. Saha et al. / Electrical Power and Energy Systems 33 (2011) 887900 895
With respect to the accelerating torque, as the speed initially
decreases, causing a decrease in the electrical torque, the rotor
shaft accelerates (T
a
= T
m
T
e
), but when the speed recovers to
zero change, the acceleration reduces to zero too, as expected.
The gures also show clearly that, due to the exciter action, the
terminal voltage is increased by the required amount of 5%.
When the fault is applied, the same chain of events takes place,
only in this case more profoundly.
5.2. Case study II: fault on both line
In this study, we consider an initially unperturbed system (i.e.
no changes in the reference inputs and no fault). Then we affect
a 5% step increase in reference voltage V
REF
at 2 s. This is followed
by a 3 phase balanced fault at 10 s in line 1 at 30% length of the
transmission line from the generator end and run the simulation
till 30 s. After 30 s we inject a fault in line 2 at 80% length of the
line from the generator end. Output responses are given in Figs.
68.
Each of the output responses are subdivided into three gures,
top one shows the overall response, middle one shows rst 17 s
and the bottom one shows the response between t = 16 s and
t = 60 s. The responses up to 30 s are same as the responses in study
case I.
After 30 s when fault takes place in line 2, the generator connec-
tion to the innite bus is completely severed. As a result the load
z
1
and its estimate (dotted)
0 5 10 15 20 25 30 35 40
z
2
and its estimate (dotted)
0 5 10 15 20 25 30 35 40
0 5 10 15 20 25 30 35 40
z
5
and its estimate (dotted)
0 5 10 15 20 25 30 35 40
z
4
and its estimate (dotted)
-20
-10
0
10
-0.5
0
0.5
0 5 10 15 20 25 30 35 40
z
3
and its estimate (dotted)
-4
-2
0
2
-10
-5
0
5
-1
0
1
2
Fig. 13. The response of the outputs of the augmented system and the corresponding observer outputs (dotted) when P
REF
is increased.
896 S. Saha et al. / Electrical Power and Energy Systems 33 (2011) 887900
angle continues slipping back at a constant rate and the speed set-
tles to a new steady state, after the transient period. The terminal
voltage undergoes a large fall immediately after the fault, as shown
in Fig. 8, but it then recovers to 5% above the initial condition due
to the action of the voltage regulator (exciter).
6. Fault detection using a sliding mode observer
Conventional fault detection algorithms are typically based on
observer theory, where the error residual between the observer
and the plant is used as an indicator for faults. Example of residual
generation algorithms incorporating unknown input observer the-
ory are found in [810], and incorporating H
1
are found in [11,12].
However, recent approaches such as sliding mode observer (SMO)
theory [13] have proven to be quite effective in fault detection in
systems with unknown inputs such as faults.
In the following we rst present a brief outline of the theory of
sliding mode observer. Readers who are not familiar with control
theory may skip this section and the next section move onto Sec-
tion 8.
For convenience let us recall the SMIB state-space equations
_ x Ax Bu E
x
f
y Cx E
y
f
45
where x 2 R
n14
; y 2 R
p5
; u 2 R
m2
; f 2 R
q1
. Since E
y
has full col-
umn rank, there exists an orthogonal matrix T
r
2 R
pp
such that
T
r
E
y

0
E
y2
_ _
; T
r
y
y
1
y
2
_ _
T
r
C
C
1
C
2
_ _
46
where E
y2
is a square and invertible matrix, and C
1
and C
2
are gen-
eral matrices with no particular structure.
Let y
f
2 R
q
be the output of a user dened stable lter driven by
y
2
:
_ y
f
A
f
y
f
A
2
y
2
A
f
y
f
A
f
C
2
x A
f
E
y2
f 47
Combine (45) and (47) to get the following augmented state-space
system of order n : n q
_ x
_ y
f
_ _

A 0
A
f
C
2
A
f
_ _
..
A
x
y
f
_ _
x

B
0
_ _
B
u
E
x
A
f
E
y2
_ _
..
Ex
f
y
1
y
f
_ _
z

C
1
0
0 I
q
_ _
..
C
x
y
f
_ _ 48
Dene z :
y
1
y
f
_ _
; z 2 R
pq6
(It is clear that z is a measurable
signal.)
Theorem 1. If the following conditions are satised
(i) rankCE
x
rankE
x
; and
(ii) the zeros of A; E
x
; C (if any) are stable,
then for the state-space model described by (48) there exists a change
of coordinates such that the triple A; E
x
; C can be re-written as:
A
A
11
A
12
A
21
A
22
_ _
; E
x

0
M
2
_ _
; C 0 T ; M
2

0
M
o
_ _
49
where A
11
2 R
np np
, matrix T 2 R
pp
is orthogonal, M
2
2 R
pq
, and
M
o
2 R
qq
is invertible. Any unobservable modes of (A
11
, A
21
) are the
invariant zeros of A; E
x
; C
_ _
and are stable. Full proof is given in [14].
By Theorem 1, a sliding mode observer of the structure shown
below can be designed for the system (48) to reconstruct the fault
f as well as the entire state and output variables.
_ w Aw Bu G
l
e
y
G
n
m 50
where e
y
Cw z and m is a nonlinear switching term dened by
m q
e
y
ke
y
k
; e
y
0; q 2 R

51
The matrices G
l
; G
n
are to be designed; in particular G
n
has the
structure G
n

LT
T
T
T
_ _
P
1
o
, where P
o
2 R
pp
is a symmetric positive
denite matrix, and L
o
2 R
nppq
. It has been proven in [15] that a
sliding motion will take place in nite time on the surface
S
e
= {e
y
= 0} if the following conditions hold:
Condition (i): there exists a matrix P with the structure
P
P
11
P
11
L
L
T
P
11
T
T
P
o
T L
T
P
11
L
_ _
> 0; P
11
2 R
np np
;
that satises the following inequality
PA G
l
C A G
l
C
T
P < 0 52
Condition (ii): The scalar q in the switching function m satises
the following inequality
q > kP
o
TM
2
k kf k:
When sliding motion has been achieved, the fault can be recon-
structed from the following signal:
0 10 20 30 40
The fault
0
0.2
0.4
0.6
0.8
1
0 10 20 30 40
The fault reconstruction
0
0.2
0.4
0.6
0.8
1
Fig. 14. The fault (left) and its reconstruction (right) when P
REF
is increased.
S. Saha et al. / Electrical Power and Energy Systems 33 (2011) 887900 897
f
e
: 0 M
1
o
_
P
1
o
m
eq
where m
eq
is a version of m required to achieve and maintain sliding
motion. The signal m
eq
is computable online by replacing m with m
eq
where
m
eq
q
e
y
ke
y
k c
53
and c is a small positive scalar that governs the degree of accuracy
of m
eq
.
7. Observer design
In this section, the theory presented in Section 6 is used to
design a SMO suitable for the power developed in Section 4. We
then demonstrate, through extensive simulations, the designed
sliding mode observers ability to detect faults on line and in real
time.
Using the data given in Appendix C, then for the given value of
E
y
, a suitable value for T
r
was found to be
T
r

1 0 0 0 0
0 1 0 0 0
0 0 1 0 1:1194
0 0 0 1 0
0 0 0 0 1
_

_
_

_
54
This transformation matrix resulted in the output equation being
transformed to the structure shown in Eq. (46). Let us partition y
and z from (45) and (48) as follows:
y
y
1
y
2
:
y
p
_

_
_

_
; z
z
1
z
2
:
z
p
_

_
_

_
Then from (46)(48) and the using T
r
from (54) and choosing A
f
= 1
we have
z
1
y
1
; z
2
y
2
; z
3
y
3
1:1194y
5
; z
4
y
4
;
_
z
5
z
5
y
5
Implementing the design method outlined above, and choosing sca-
lar parameters q, c associated with the nonlinear switching term
(53) as 1500 and 0.0001, respectively, the following observer gains
are found.
G
l
G
n

0:0076 0:0344 0:2477 0:0531 0:0149
0:0148 0:0318 0:0401 0:0247 0:0244
0:0112 0:0403 0:0456 0:0334 0:0191
0:0083 0:0476 0:7441 0:6949 0:0067
0:0054 0:0043 0:0040 0:0036 0:0038
0:0069 0:0107 0:0089 0:0101 0:0055
0:0131 0:0074 0:0021 0:0095 0:0122
0:0074 0:6588 0:9079 0:3008 0:0013
0:0137 0:0365 2:1023 1:6759 0:0163
0:0132 0:0084 0:0291 0:0476 0:0154
0:2084 0:0388 1:0412 1:0056 0:2118
0:0002 0:0018 0:0017 0:0017 0
0:0001 0:0011 0:0013 0:0009 0:0005
0:0004 0:0013 0:0059 0:0050 0:0080
0:0122 0:0013 0:0128 0:0058 0:0371
_

_
_

_
8. Simulation results
In this section we investigate the performance of the SMO on
the sample power system described in (45). The following scenario
is simulated: The power system initially runs unperturbed (steady-
state) condition for 5 s. Then the excitation system reference com-
mand, V
REF
, is increased by 5% at time t = 5 s. This is followed by
applying a solid three-phase line-to-ground fault half way along
line 2 at time t = 25 s.
The simulation results are shown in Figs. 914, where the fol-
lowing initial conditions are used: The power system states state
vector, x, was assigned the value calculated from the initialisation
process, which is
x = [0.9501, 0.2311, 0.6068, 0.4860, 0.8913, 0.7621, 0.4565, 0.0185, -
0.8214, 0.4447, 0.6154, 0.7919, 0.9218, 0.7382]. The observer was
assumed to have zero initial conditions (this is normally the situa-
tion in practice when the observer is rst connected to the power
system and switched on).
Fig. 9 shows the responses of the output variables, which are ex-
actly the as those shown in Fig. 5 (please refer to Section 5 for com-
mentary on the behaviour of these variables). Fig. 10 shows the
outputs of the augmented system as well as the outputs of the fault
lter (superimposed with dotted lines). Visually, there is no differ-
ence between the two sets of responses despite the presence of the
fault and the mismatch in the initial conditions (though upon
zooming in, it can be seen that there is an initial mismatch be-
tween the augmented system outputs and the observer outputs,
but perfect convergence takes place after 0.05 s).
Fig. 10 also shows that the fault lter is also a state estimator
and is able to reconstruct exactly the output variables. Further-
more the gure shows a very important characteristic of the fault
lter; that is it is only sensitive to faults and no any other internal
or external inputs.
Fig. 11 shows the performance of the fault lter, which demon-
strates quite clearly its ability to detect and reconstruct the fault in
near real time despite the fact that it starts from a different set of
initial conditions. In practice, however, the fault lter is continu-
ously switched on and therefore is able to detect faults in real time
whenever faults occur. Note that from the period up to the applica-
tion of fault, the fault lter does not respond despite the fact that a
change in the external input is affected.
Figs. 1214 show the case where the reference power, P
REF
, is in-
creased. Here again the fault reconstruction takes place almost in
real time, despite the difference in the initial conditions.
9. Conclusions
This paper presents a new fault dependent model for a single
machine connected to an innite bus through double transmission
lines. The model is simulated for faults occurring in either of the
transmission lines and in both and the results are discussed in de-
tail, where it is explained how the model behaves in the manor it
does in real life.
Then a fault detection lter (locator) is designed using sliding
mode observer theory. The lter is then tested, through extensive
simulations, on the power system where disturbances and faults
are applied. It is demonstrated quite clearly that the designed fault
detector is only sensitive to faults and not to any other external
disturbance, such changes in the reference inputs, load demands,
noise, etc. furthermore the lter is also capable of acting as a state
estimator where the states, and outputs can be reconstructed in
their entireties.
Extension of this research is now underway to the multi-
machine case and also to the case where asymmetrical faults occur
in the transmission network.
898 S. Saha et al. / Electrical Power and Energy Systems 33 (2011) 887900
Appendix A. State equations parameters
Combining the state equations of the synchronous machine, ex-
citer and governor results in the following state equation.
_ x A
x
x A
i
I Bu Ef
where the current I is dened as follows:
I I
T
x
0
T
8
_
; I
x
i
d
i
F
i
D
i
q
i
G
i
Q
_
T
The current I
x
can be expressed in terms of the states, through the
following ux linkagecurrent relationship.
I
x
L
1
x
w
; L
L
d
L
ad
L
ad
0 0 0
L
ad
L
F
L
ad
0 0 0
L
ad
L
ad
L
D
0 0 0
0 0 0 L
q
L
aq
L
aq
0 0 0 L
aq
L
G
L
aq
0 0 0 L
aq
L
aq
L
Q
_

_
_

_
As a result, the state equation can now be written as
_ x Ax Bu E
x
f ; A A
x
A
i
L
inverse
:
The matrix A is 14 14 real matrix with the nonzero elements
being
A
x
1; 4 x
b
1
_
d

; A
x
1; 7 x
b
jV
B
j cos d

A
x
1; 8 x
b
w

q
A
x
2; 11
x
b
r
F
L
ad
A
x
4; 1 x
b
1
_
d

A
x
4; 7 x
b
jV
B
j sind

; A
x
1; 8 x
b
w

d
A
x
7; 8 1
A
x
8; 1
x
b
2H
i

q
A
x
8; 4
x
b
2H
i

d
A
x
8; 8
x
b
2H
k
d

T

m
1
_
d

_ _
A
x
9; 7
K
R
T
R
a
c
jV
B
j v

d
cos d

q
sind

_ _
A
x
9; 9
1
T
R
; A
x
10; 9
K
A
K
F
T
A
T
F
A
x
10; 10
1
T
F

K
A
K
F
T
A
T
F
_ _
A
x
10; 11
K
F
T
A
T
F
A
x
11; 9
K
A
T
A
; A
x
11; 10
K
A
T
A
; A
x
11; 11
1
T
A
A
x
12; 8
1
x
b
T
SR
R
Gov
; A
x
12; 12
1
T
SR
A
x
13; 12
1
T
SM
; A
x
13; 13
1
T
SM
A
x
14; 13
1
T
ST
; A
x
14; 14
1
T
ST
The matrix A
i
is also 14 14 real matrix with the following nonzero
entries
A
i
1; 1 x
b
r
R
2
_ _
; A
i
1; 4
x
b
X
2
A
i
2; 2 x
b
r
F
; A
i
3; 3 x
b
r
D
A
i
4; 1
x
b
X
2
; A
i
4; 4 x
b
r
R
2
_ _
A
i
5; 5 x
b
r
G
; A
i
6; 6 x
b
r
Q
A
i
8; 1
x
b
2H
w

q
; A
i
8; 4
x
b
2H
w

d
A
i
9; 1
K
R
T
R
a
c
v

d
R
2

K
R
T
R
a
c
v

q
X
2
_ _
A
i
9; 4
K
R
T
R
a
c
v

q
R
2

K
R
T
R
a
c
v

d
X
2
_ _
where
a
c

1

v
2
d
v
2
q
_
v
0
d
and v
0
q
are initial values of v
d
and v
q
, respectively.
The matrix B is 14 2 real matrix with the following nonzero
entries
B10; 1
K
A
K
F
T
A
T
F
; B10; 1
K
A
T
A
;
B12; 2
1
T
SR
E
x
is a 14 2 matrix with the following nonzero elements.
E
x
1; 1 b
1
1x
b
jV
B
j sind

b
1

1
2
_ _
x
b
Ri

d
b
1

1
2
_ _
x
b
Xi

q
E
x
4; 1 b
1
1x
b
jV
B
j cos d

b
1

1
2
_ _
x
b
Ri

q
b
1

1
2
_ _
x
b
Xi

d
E
x
9; 1
b
1
1jV
B
j sind
K
R
T
R
a
c
v

K
R
T
R
a
c
v

d
b
1

1
2
_ _
Ri

d

K
R
T
R
a
c
v
0
d
b
1

1
2
_ _
Xi
0
q
_
_
_
_

b
1
1jV
B
j cos d
K
R
T
R
a
c
v

K
R
T
R
a
c
v

q
b
1

1
2
_ _
Ri

q

K
R
T
R
a
c
v

q
b
1

1
2
_ _
Xi

d
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
E
x
1; 2 b
2
1x
b
jV
B
j sind

b
2

1
2
_ _
x
b
Ri

d
b
2

1
2
_ _
x
b
Xi

q
E
x
4; 2 b
2
1x
b
jV
B
j cos d

b
2

1
2
_ _
x
b
Ri

q
b
2

1
2
_ _
x
b
Xi

d
E
x
9; 2
b
2
1jV
B
j sind
K
R
T
R
a
c
v

K
R
T
R
a
c
v

d
b
2

1
2
_ _
Ri

d

K
R
T
R
a
c
v
0
d
b
2

1
2
_ _
Xi
0
q
_
_
_
_

b
1
1jV
B
j cos d
K
R
T
R
a
c
v

K
R
T
R
a
c
v

q
b
2

1
2
_ _
Ri

q

K
R
T
R
a
c
v

q
b
2

1
2
_ _
Xi

d
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
Appendix B. State equations parameters
The output equation is dened in (44) as
y d
_
d V
N
T
a
P
N

T
The rst two are dened as state variables. The terminal voltage is
derived in Eq. (38). The accelerating torque T
a
is dened as
S. Saha et al. / Electrical Power and Energy Systems 33 (2011) 887900 899
T
a
= T
m
T
e
, where T
e
and T
m
are given in (8) and (15), respectively.
The terminal power of the generator, P
N
is written as
P
N
V

N
I
N
I

N
V
N
where I
0
N
and V
0
N
are initial values of terminal voltage and current,
respectively, and
I
N

i
d
i

d
i
q
i

q
_ _

i
2
d
i
2
q
_ and V
N

v
d
v

d
v
q
v

v
2
d
v
2
q
_
Combining these expressions, yields the following output equation
y Cx E
y
f
where C = (C
x
+ C
i
L
inverse
), C
x
is a 5 14 matrix with the following
nonzero elements:
C
x
1; 7 1; C
x
2; 8 1
C
x
3; 7 a
c
v

d
jV
B
j cosd

a
c
v

q
jV
B
j sind

C
x
4; 1 i

q
; C
x
4; 4 i

d
C
x
4; 8 T

m
=1
_
d

; C
x
4; 14 1=1
_
d

C
x
5; 7 I

N
a
c
v

d
jV
B
jCosd

a
1
v

c
jV
B
jSind

_ _
C
i
is a 5 14 matrix with the following nonzero elements
C
i
3; 1
1
2
a
c
v

d
R
1
2
a
c
v

q
X; C
i
3; 4
1
2
a
c
v

d
X
1
2
a
c
v

q
R
Ci4; 1 w

q
; Ci4; 4 w

d
Ci5; 1
1
2
a
c
v

d
R
1
2
a
c
v

q
X
_ _
I

N
V

N
bi

d
Ci5; 4
1
2
a
c
v

d
X
1
2
a
c
v

q
R
_ _
I

N
V

N
bI

q
where
b
1

i
0
2
d
i
2
q
_
E
y
is a 5 2 matrix with the following nonzero elements
Ey3; 1
a
c
v

d
b
1
1V
B
sind

a
c
v

d
b
1

1
2
_ _
Ri

d
a
c
v

d
b
1

1
2
_ _
Xi

q
a
c
v

q
b
1
1V
B
cos d

a
c
v

q
b
1

1
2
_ _
Ri

q
a
c
v

q
b
1

1
2
_ _
Xi

d
_
_
_
_
_
_
Ey5; 1
a
c
v

d
b
1
1V
B
sind
0
a
c
v

d
b
1

1
2
_ _
Ri

a
c
v

d
b
1

1
2
_ _
Xi

q
a
c
v

q
b
1
1V
B
cos d
0

a
c
v

q
b
1

1
2
_ _
Ri

q
a
c
v

q
b
1

1
2
_ _
Xi

d
_
_
_
_
_
_
_
_
I

N
Ey3; 2
a
c
v

d
b
2
1V
B
sind

a
c
v

d
b
2

1
2
_ _
Ri

d
a
c
v

d
b
2

1
2
_ _
Xi

q
a
c
v

q
b
2
1V
B
cos d

a
c
v

q
b
2

1
2
_ _
Ri

q
a
c
v

q
b
2

1
2
_ _
Xi

d
_
_
_
_
_
_
Ey5; 2
a
c
v

d
b
2
1V
B
sind
0
a
c
v

d
b
2

1
2
_ _
Ri

a
c
v

d
b
2

1
2
_ _
Xi

q
a
c
v

q
b
2
1V
B
cos d
0

a
c
v

q
b
2

1
2
_ _
Ri

q
a
c
v

q
b
2

1
2
_ _
Xi

d
_
_
_
_
_
_
_
_
I

N
Appendix C. Data
L
d
0:2768; L
ad
0:2497; L
aq
0:2422; L
q
0:2692
L
f
0:2795; L
G
0:3804; L
Q
0:2647; R 0:0200
X 0:2000; kd 1; H 2:3700 KW s=KVA; x
b
2pf rad=s
K
R
1; T
R
0:0001 s; K
A
400; T
A
0:0500 s
K
F
0:0400; T
F
1 s; R
gov
0:0500; T
SR
0:1000 s
T
SM
0:2500 s; T
ST
1 s
References
[1] Aldeen M, Sharma R. Robust detection of faults in frequency control loops. IEEE
Trans Power Syst 2007;22:41322.
[2] Crusca F, Aldeen M. Fault detection and identication of power systems. In:
IASTED-intelligent systems & control (ISC03). Salzburg, Austria: IASTED; 2003.
p. 1838.
[3] IEEE, Recommended practice for excitation system models for power system
stability studies. IEEE transactions on PAS, vol. IEEE Working Group on
Excitation Systems, IEEE Std 421.5-1992; 1992.
[4] IEEE, MW response of fossil fuelled units, IEEE transactions on PAS, vol.
Working group on power plant response to load changes; 1972. p. 45563.
[5] Kundur P, Balu NJ, Lauby MG. Power system stability and control. New
York: McGraw-Hill; 1994.
[6] Anderson PM, Agrawal BL, Van Ness JE, and IEEE Power Engineering Society.
Subsynchronous resonance in power systems. New York: IEEE Press; 1990.
[7] Anderson PM, Fouad AA, and Institute of Electrical and Electronics Engineers
(US). Power system control and stability. Piscataway, NJ: IEEE Press; 1993.
[8] Hou M, Muller PC. Design of observers for linear systems with unknown
inputs. IEEE Trans Autom Control 1992;37:8715.
[9] Wang SH, Davison EJ, Dorato P. Observing the states of systems with
unmeasurable disturbances. IEEE Trans Autom Control 1975;20:7167.
[10] Darouach M, Zasadzinski M, Xu SJ. Full-order observers for linear systems with
unknown inputs. IEEE Trans Autom Control 1994;39:6069.
[11] Hou M, Patton RJ. An LMI approach to H_innity fault detection observers, in
control 96. UKACC international conference on (Conf. Publ. No. 427), vol. 1;
1996. p. 30510.
[12] Ibaraki S, Suryanarayanan S, Tomizuka M. H-innity optimization of
Luenberger state observers and its application to fault detection lter
design. Presented at proceedings of the 40th IEEE conference on decision
and control; 2001.
[13] Tan CP, Edwards C. Sliding mode observers for detection and reconstruction of
sensor faults. Automatica 2002;38:181521.
[14] Edwards C, Spurgeon SK. On the development of discontinuous observers. Int J
Control 1994;59:121129.
[15] Tan CP, Edwards C. Sliding mode observers for robust detection and
reconstruction of actuator and sensor faults. Int J Robust Nonlinear Control
2003;13:44363.
[16] Eisa A, Kumar k. Accurate one end fault location for overhead transmission
lines in interconnected power systems. Int J Electr Power Energy Syst
2010;32:3839.
[17] Samantary SR, Dash PK, Upadhyay SK. Adaptive Kalman lter and neural
network based high impedance fault detection in power distribution
networks. Int J Electr Power Energy Syst 2009;31:16772.
[18] Bhowmik PS, Purkait P, Bhattacharya K. A novel wavelet transform aided
neural network based transmission line fault analysis method. Int J Electr
Power Energy Syst 2009;31:2139.
900 S. Saha et al. / Electrical Power and Energy Systems 33 (2011) 887900