13-08-2014, 04:14 PM
RROJECT REPORT ON LONG TRANSMISSION LINE PROTECTION USING ARTIFICIAL NEURAL NETWORK
[attachment=66626]
Abstract
This paper proposes an effective fault detection
technique in distance relaying using ANN based Euclidean
distance function based on discrete current samples. The
original signal under fault consist two parts namely normal
and disturbance part. The fault detection is easily achieved
as the disturbance part of the signal produces an irregular
shape compared to the shape produced from the normal
part of the signal. By selecting suitable threshold value the
starting point of irregular part can be found. The proposed
algorithm is so effective and easy when compared to other
algorithms. Results are carried out in
MATLAB/SIMULINK softwar
INTRODUCTION
Power system protection plays an important role in present
days. To protect power system proper operation of relays,
circuit breakers are necessary. The inception of a fault in the
power system, the voltage and current waveforms deviate from
the normal sinusoid. Fault detection algorithms in a relay (for
example, a distance relay) employ either of these signals to
discriminate the faulty situation from the normal state. The
sample-to-sample comparison of the current (voltage) signal is
a straightforward approach for fault detection. Another simple
approach is the comparison of the present current sample with
the corresponding value before one or two cycles’ faults on
transmission lines need to be detected, classified, and cleared as
fast as possible. In power transmission-line protection, fault
detection and fault classification are the two most important
items which need to be addressed in a reliable and accurate
manner. Fault detection plays an important role in distance
relaying as it discriminates the normal state from the fault and
activates the main relay algorithm. The overall performance of
a relay depends very much on the speed and accuracy of the
fault detection technique
EXISTED FAULT DETECTION METHODS
Sample to Sample Comparison technique:
In this method the algorithm is written based on sample to
sample comparison of a current signal or a voltage signal.
Under normal case the difference between samples to next
sample is same. If any disturbance occur in the system leads to
the deviation in the difference value. By fixing a threshold
value, these situations can be detected.
Mathematically
CUMULATIVE SUM APPROACH
principle and are therefore sensitive to noise or spikes in the
signal.. The proposed CUSUM method for fault detection uses
the current samples of any phase and prepares two
complementary signals as
Sk (1) = Sk (5)
Sk (2) = - Sk (6)
Using the aforementioned two signals, the two-sided
CUMSUM test is expressed as
gk (1) = max (gk – 1(1) +Sk(1) –v, 0) (7)
gk (1) = max (gk – 1(1) +Sk(1) –v, 0) (8)
Where v represents the test statistics and is the drift
parameter in it.
A fault is registered if
gk (1) > h or gk (2) > h (9)
Where h is an arbitrary constant and which should be ideally
zero.
PROPOSED METHOD
In this paper ANN based Euclidean distance function based
discrete current sample based algorithm is proposed and the
effectiveness of the proposed algorithm is verified under
different situations namely Different types of fault (L-G, LL-G,
LL and LLL-G),Different fault resistances and fault inception
angles, At different lengths of the line, Capacitive switching,
Load change (increase and decrease), Spike.
The proposed method depends on voltage and curre
Fault with different resistances
Impedance based classification is one of the type of
classification for faults. Generally faults having low impedance
and they can be easily detected. But faults with high impedance
are cannot be detected by all algorithms. The proposed
algorithm detects every type faults with high impedance that are
occur on transmission line network. In this paper invistigations
are carried out by giving all high impedance faults randomly
with different fault impedances.some of the proposed algorithms
in earlier days can detect high impedance faults but there exsit
some limitations and it is observed that all type are detected by
the present proposed algorithm in this paper. Simulations are
carried out and results are shown in figures
CONCLUSION
A simple approach for the fault detection of power
system relaying is proposed. The relative performance of the
algorithm is considered with signals containing noise, spike,
frequency deviation, load change, and faults at different
situations of the power system. The method is found to be a
better tool to be used with power system relays and can be tried
for other applications such as faulty phase selection, fault
classification, and in power-quality disturbance detection.
[attachment=66626]
Abstract
This paper proposes an effective fault detection
technique in distance relaying using ANN based Euclidean
distance function based on discrete current samples. The
original signal under fault consist two parts namely normal
and disturbance part. The fault detection is easily achieved
as the disturbance part of the signal produces an irregular
shape compared to the shape produced from the normal
part of the signal. By selecting suitable threshold value the
starting point of irregular part can be found. The proposed
algorithm is so effective and easy when compared to other
algorithms. Results are carried out in
MATLAB/SIMULINK softwar
INTRODUCTION
Power system protection plays an important role in present
days. To protect power system proper operation of relays,
circuit breakers are necessary. The inception of a fault in the
power system, the voltage and current waveforms deviate from
the normal sinusoid. Fault detection algorithms in a relay (for
example, a distance relay) employ either of these signals to
discriminate the faulty situation from the normal state. The
sample-to-sample comparison of the current (voltage) signal is
a straightforward approach for fault detection. Another simple
approach is the comparison of the present current sample with
the corresponding value before one or two cycles’ faults on
transmission lines need to be detected, classified, and cleared as
fast as possible. In power transmission-line protection, fault
detection and fault classification are the two most important
items which need to be addressed in a reliable and accurate
manner. Fault detection plays an important role in distance
relaying as it discriminates the normal state from the fault and
activates the main relay algorithm. The overall performance of
a relay depends very much on the speed and accuracy of the
fault detection technique
EXISTED FAULT DETECTION METHODS
Sample to Sample Comparison technique:
In this method the algorithm is written based on sample to
sample comparison of a current signal or a voltage signal.
Under normal case the difference between samples to next
sample is same. If any disturbance occur in the system leads to
the deviation in the difference value. By fixing a threshold
value, these situations can be detected.
Mathematically
CUMULATIVE SUM APPROACH
principle and are therefore sensitive to noise or spikes in the
signal.. The proposed CUSUM method for fault detection uses
the current samples of any phase and prepares two
complementary signals as
Sk (1) = Sk (5)
Sk (2) = - Sk (6)
Using the aforementioned two signals, the two-sided
CUMSUM test is expressed as
gk (1) = max (gk – 1(1) +Sk(1) –v, 0) (7)
gk (1) = max (gk – 1(1) +Sk(1) –v, 0) (8)
Where v represents the test statistics and is the drift
parameter in it.
A fault is registered if
gk (1) > h or gk (2) > h (9)
Where h is an arbitrary constant and which should be ideally
zero.
PROPOSED METHOD
In this paper ANN based Euclidean distance function based
discrete current sample based algorithm is proposed and the
effectiveness of the proposed algorithm is verified under
different situations namely Different types of fault (L-G, LL-G,
LL and LLL-G),Different fault resistances and fault inception
angles, At different lengths of the line, Capacitive switching,
Load change (increase and decrease), Spike.
The proposed method depends on voltage and curre
Fault with different resistances
Impedance based classification is one of the type of
classification for faults. Generally faults having low impedance
and they can be easily detected. But faults with high impedance
are cannot be detected by all algorithms. The proposed
algorithm detects every type faults with high impedance that are
occur on transmission line network. In this paper invistigations
are carried out by giving all high impedance faults randomly
with different fault impedances.some of the proposed algorithms
in earlier days can detect high impedance faults but there exsit
some limitations and it is observed that all type are detected by
the present proposed algorithm in this paper. Simulations are
carried out and results are shown in figures
CONCLUSION
A simple approach for the fault detection of power
system relaying is proposed. The relative performance of the
algorithm is considered with signals containing noise, spike,
frequency deviation, load change, and faults at different
situations of the power system. The method is found to be a
better tool to be used with power system relays and can be tried
for other applications such as faulty phase selection, fault
classification, and in power-quality disturbance detection.