To identify and simulate conventional type of disturbance on the overhead transmission line by using PSCAD / EMTDC software package To develop mathematical model for various type of disturbance on overhead transmission line. To develop a smart algorithm for fault detection using Artificial Neural Network (ANN) and Particle Swarm Optimization (PSO).

1. UTMUNIVERSITI TEKNOLOGI MALAYSIA
FAULT DETECTION
ON OVERHEAD TRANSMISSION LINE
USING ARTIFICIAL NEURAL NETWORK AND
PARTICLE SWARM OPTIMIZATION
By
MAKMUR SAINI
SUPERVISED BY
PROF.IR.DR.ABDULLAH ASUHAIMI BIN MOHD ZIN
CO SUPERVISOR BY
PROF.DR.MOHD WAZIR BIN MUSTAFA

2. Table Of Content
OBJECTIVES
SCOPE OF THE RESEARCH
The types of fault that will be simulated
Overview Short Circuit Fault Analysis
Research Methodology
PRELIMINARY RESULT
CONCLUSION

3. OBJECTIVES
1. To identify and simulate conventional type of
disturbance on the overhead transmission line by
using PSCAD / EMTDC software package
2. To develop mathematical model for various type of
disturbance on overhead transmission line.
3. To develop a smart algorithm for fault detection
using Artificial Neural Network (ANN) and Particle
Swarm Optimization (PSO).

4. SCOPE OF THE RESEARCH
1. Identification and simulation of various of
disturbance on overhead transmission line by
using PSCAD/EMTDC software. Version 4.2.0
2. Preparing suitable mathematical model for voltage
and current signals of the above disturbances.
3. Development of the proposed smart algorithm by
using Artificial Neural Network (ANN) and
Particle Swarm Optimization (PSO) method in
fault detection of overhead transmission line.

5. The types of fault that will be simulated
The single line to ground fault
The line to line fault
The double line to ground fault
Three phases of to ground fault
 The lighting Strike fault

6. Overview Short Circuit Fault Analysis
Transient short circuit on the transmission line can be
simplified with certain assumptions based on the
following stages:
The line is fed from a constant voltage source
Short circuit takes place when the line is
unloaded
Line capacitance is negligible, and the line can
be represented by a lumped RL series

7. Overview Short Circuit Fault Analysis
Figure Transmission Line Model and Waveform of Short circuit current

8. Overview Short Circuit Fault Analysis
dcactot iii +=
dcactot iii +=
)sin( φα++= wtIIac
t
L
R
dc etII
)(
)sin(
−
+−= αω
])sin())[sin(
)( t
L
R
tot ettII
−
+−−+= αωθαω

9. Research Methodology
 Fault detection is proposed by creating
a simulation current and voltage signals
at several fault conditions that obtained
through simulation using PSCAD/
EMTDC.
The waveforms obtained in simulation
PSCAD will be trained using ANN - PSO
method with the Matlab program

10. Research Methodology
The results form the signal currents and
voltages are similar when compared to
results obtained from the pattern of
training ANN-PSO
Expected result to generate a simulation
model of fault detection and faults on
overhead transmission line path by using
ANN-PSO.

11. BLOK DIAGRAM OF THE RISET

12. Flowchart for Learning the ANN using PSO algorithm.
.
ididid
idgdidididid
VXX
XPrcXPrcWxVV
+=
−+−+= )()( 2211

13. ALOGARITHMS FAULT DETECTION

14. ALOGARITHMS FAULT DETECTION

15. PROGRESSRESULT
The study was conducted using of
PSCAD/EMTDC that generate current , voltage
wave signal and Mathematical Model. Below
are the 5 types of fault
 The line to ground fault
 The line to line fault
 The line-line to ground fault
 The three phase to ground fault
 The lightning strike fault

16. PROGRESSRESULT

17.
Voltage Waveform Signal Fault Line to Ground ( LG )

18.
Current Waveform Signal Fault Line to Ground ( LG )

19. Model Mathematic Voltage and Current Signal Original
Fault Line to Ground (LG)
kVtV xa )
12
13
cos(8.53)(1 πω+=
kAetI
t
L
R
xa ])
3
1
sin()
3
1
[sin(995.1
)(
)(1
−
−−−= ππω
kVtV xb )
12
15
cos(4.145)(1 πω−=
kVtV xc )
6
11
cos(9.133)(1 πω−=
kAetI
t
L
R
xc ])
4
3
sin()
4
3
[sin(558.0
)(
)(1
−
−−+= ππω
kAetI
t
L
R
xb ])
3
1
sin()
3
1
[sin(345.0
)(
)(1
−
−−= ππω

20. Voltage and Current, Sampling the Signal for N sample per cycle
Fault Line to Ground (LG)
kV
N
n
V na )
12
13
60
cos(8.53)(1 πω +=
kAe
N
n
I N
n
na ])
3
1
sin()
3
1
60
[sin(995.1
)
60
(
)(1
τ
ππ
ω −
−−−=
kV
N
n
V nb )
12
15
60
cos(4.145)(1 πω −=
kV
N
n
V nc )
6
11
60
cos(9.133)(1 πω −=
kAe
N
n
I
t
N
n
nc ])
4
3
sin()
4
3
60
[sin(558.0
)
60
(
)(1
τ
ππ
ω −
−−+=
kAe
N
n
I N
n
nb ])
3
1
sin()
3
1
60
[sin(345.0
)
60
(
)(1
τ
ππ
ω −
−−=
L
R
=τ
N = Sample per cycle of Data
n = 1 ,2, ……….N-1

21. Voltage and Current Fourier Transform Of this Sequence ,
Fault Line to Ground (LG)
∑
−
=
−
=
1
0
)2(
)(1)(1
N
n
N
nkj
naka eVV
π
∑
−
=
−
=
1
0
)2(
)(1)(1
N
n
N
nkj
nbkb eVV
π
∑
−
=
−
=
1
0
)2(
)(1)(1
N
n
N
nkj
kckc eVV
π
∑
−
=
−
=
1
0
)2(
)(1)(1
N
n
N
nkj
kaka eII
π
∑
−
=
−
=
1
0
)2(
)(1)(1
N
n
N
nkj
nbkb eII
π
∑
−
=
−
=
1
0
)2(
)(1)(1
N
n
N
nkj
nckc eII
π

22.
Voltage Waveform Signal Fault Line to Line Ground ( LLG )

23.
Current Waveform Signal Fault Line to Line Ground (L LG )

24. Model Mathematic Voltage and Current Signal Original
Fault Line to Line Ground (LLG)
kVtV xa )cos(8.53)(1 ω=
kAetI
t
L
R
xa ])
6
1
sin()
6
1
[sin(076.2
)(
)(1
−
−−−= ππω
kVtV xb )
6
5
cos(4.138)(1 πω +=
kVtV xc )
2
1
cos(2.48)(1 πω −=
kAetI
t
L
R
xc ])
4
3
sin()
4
1
[sin(534.1
)(
)(1
−
−−+= ππω
kAetI
t
L
R
xb ])
3
2
sin()
3
2
[sin(496.0
)(
)(1
−
−−+= ππω

25. Voltage and Current, Sampling the Signal for N sample per cycle
Fault Line-Line to Ground (LG)
kV
N
n
V na )
60
cos(8.53)(1
ω=
kAe
N
n
I N
n
na ])
6
1
sin()
6
1
60
[sin(076.2
)
60
(
)(1
τ
ππ
ω −
−−−=
kV
N
n
V nb )
6
5
60
cos(4.138)(1 πω +=
kV
N
n
V nc )
2
1
60
cos(20.48)(1 πω −=
kAe
N
n
I
t
N
n
nc ])
4
3
sin()
4
3
60
[sin(534.1
)
60
(
)(1
τ
ππ
ω −
−−+=
kAe
N
n
I N
n
nb ])
3
2
sin()
3
2
60
[sin(496.0
)
60
(
)(1
τ
ππ
ω −
−+=
L
R
=τ
N = Sample per cycle of Data
n = 0 ,1 ,2, ……….N-1

26. Voltage and Current Fourier Transform Of this Sequence ,
Fault Line to Line Ground (LLG)
∑
−
=
−
=
1
0
)2(
)(1)(1
N
n
N
rkj
naka eVV
π
k = 0 ,1 , ………….N-1
∑
−
=
−
=
1
0
)2(
)(1)(1
N
n
N
nkj
nbkb eVV
π
∑
−
=
−
=
1
0
)2(
)(1)(1
N
n
N
nkj
nckc eVV
π
∑
−
=
−
=
1
0
)2(
)(1)(1
N
n
N
nkj
naka eII
π
∑
−
=
−
=
1
0
)2(
)(1)(1
N
n
N
nkj
nbkb eII
π
∑
−
=
−
=
1
0
)2(
)(1)(1
N
n
N
nkj
ncrc eII
π

27. Discrete Fourier Transform (DFT)
this Sequence Current (Ia) Fault Line to Ground (LG)

28. Discrete Fourier Transform (DFT)
this Sequence Current (Ib) Fault Line to Ground (LG)

29. Discrete Fourier Transform (DFT)
this Sequence Current (Ic) Fault Line to Ground (LG)

30. 30