The transmission overhead line is one of the vital elements in the power system for transmitting the electrical energy. In the transmission, the disturbances are often occurred. In the conventional algorithm, alpha and beta (mode) currents generated by Clarke’s transformation are utilized to convert the signal of Discrete Wavelet Transform (DWT) to obtain the Wavelet Transform Coefficient (WTC) and the Wavelet Coefficient Energy (WCE). This study introduces a new algorithm, called Modified Clarke for fault detection and classification using DWT and Back-Propagation Neural Network (BPNN) based on Clarke’s transformation on transmission overhead line by adding gamma current in the system. Daubechies4 (Db4) is used as a mother wavelet to decompose the high frequency components of the signal error. Simulation is performed using PSCAD / EMTDC transmission system modeling and carried out at different locations along the transmission line with different types of fault, fault resistances, fault locations and fault of the initial angle on a given power system model. The simulated fault types are in the study are the Single Line to Ground, the Line To Line, the Double Line to Ground and the Three Phases. There are four statistic methods utilized in the present study to determine the accuracy of detection and classification of faults. The result shows that the best and the worst structures of BPNN occurred on the configuration of 12-24-48-4 and 12-12-6-4, respectively. For instance, the error using Mean Square Error Method. The Error Of Clarke’s, Without Clarke’s and Modified Clarke’s are 0.05862, 0.05513 and 0.03721 which are the best, respectively, whereas, the worst are 0.06387, 0.0753 and 0.052, respectively. This indicates that the Modified Clarke’s result is in the lowest error. The method is successfully implement can be utilized in the detection and classification of fault in transmission line by utilities and power regulation in power system planning and operation.
FEA Based Level 3 Assessment of Deformed Tanks with Fluid Induced Loads
FAULT DETECTION AND CLASSIFICATION ON TRANSMISSION OVERHEAD LINE USING BPPN AND WAVELET TRANSFORMATION BASED ON CLARKE’S TRANSFORMATION
1. UTMUNIVERSITI TEKNOLOGI MALAYSIA
FAULT DETECTION AND CLASSIFICATION
ON TRANSMISSION OVERHEAD LINE
USING BPPN AND WAVELET
TRANSFORMATION BASED ON CLARKE’S
TRANSFORMATIONBy
MAKMUR SAINI
SUPERVISED BY
PROF.IR.DR.ABDULLAH ASUHAIMI BIN MOHD ZIN
CO SUPERVISOR BY
PROF.IR.DR.MOHD WAZIR BIN MUSTAFA
2. Abstract
The transmission overhead line is one of the vital elements in the power
system for transmitting the electrical energy. In the transmission, the
disturbances are often occurred. In the conventional algorithm, alpha and
beta (mode) currents generated by Clarke’s transformation are utilized to
convert the signal of Discrete Wavelet Transform (DWT) to obtain the
Wavelet Transform Coefficient (WTC) and the Wavelet Coefficient Energy
(WCE). This study introduces a new algorithm, called Modified Clarke for
fault detection and classification using DWT and Back-Propagation Neural
Network (BPNN) based on Clarke’s transformation on transmission overhead
line by adding gamma current in the system. Daubechies4 (Db4) is used as
a mother wavelet to decompose the high frequency components of the signal
error. Simulation is performed using PSCAD / EMTDC transmission system
modeling and carried out at different locations along the transmission line
with different types of fault, fault resistances, fault locations and fault of the
initial angle on a given power system model
3. Abstract
The simulated fault types are in the study are the Single Line to Ground, the
Line To Line, the Double Line to Ground and the Three Phases. There are
four statistic methods utilized in the present study to determine the accuracy
of detection and classification of faults. The result shows that the best and
the worst structures of BPNN occurred on the configuration of 12-24-48-4
and 12-12-6-4, respectively. For instance, the error using Mean Square Error
Method. The Error Of Clarke’s, Without Clarke’s and Modified Clarke’s are
0.05862, 0.05513 and 0.03721 which are the best, respectively, whereas,
the worst are 0.06387, 0.0753 and 0.052, respectively. This indicates that the
Modified Clarke’s result is in the lowest error. The method is successfully
implement can be utilized in the detection and classification of fault in
transmission line by utilities and power regulation in power system planning
and operation.
4. Introduction
The proposed approach combines the decomposition of
electromagnetic wave propagation modes, using the Clarke’s
transformation of signal processing, given by the discrete
wavelet transformation based upon the maximum signal
amplitude (WTC) 2
to determine the time intrusion. We made
extensive use of simulation software PSCAD / EMTDC which
resulted in fault of the simulation of the transient signal
transmission line parallel with the number of data points. into a
two-phase signal.
5. Introduction
For one kind of fault, this data is then transferred to
MATLAB with the help of Clarke’s transformation to
convert the three-phase signal.
The signal is then transformed into Mother Wavelet.
We manipulated several mothers wavelet such as DB4,
Sym4, Coil4 and Db8 for comparison in terms of time
and the distance estimation fault in parallel
transmission line.
.
6. . Clarke’s Transformation
Clarke's transformation, also referred to as (αβ) transformation, is
a mathematical transformation to simplify the analysis of a series
of three phases (a, b, c). It is a two-phase circuit (αβ0) stationery
and conceptually very similar to the (dqo) transformation.
= =
7. Fault Characterization in Clarke’s Transformation
1. Single line to Ground Fault (AG)
The egg line to ground fault (AG), assuming grounding resistance is zero. The instantaneous boundary
conditions are : = = 0 and = 0
then the boundary condition instantaneous are:
= 2/3 ; = 0; and = 1/3
2 Line to line Fault (AB)
The egg line to ground fault (AB), assuming grounding resistance is zero. The instantaneous boundary
conditions are : = 0 = - and = -
then the boundary condition instantaneous are:
= , = - and = 0
3 Line to line to Ground Fault (ABG)
The egg line to ground fault (ABG), assuming grounding resistance is zero. The instantaneous boundary
conditions are : = 0 , = and = = 0
then the boundary condition instantaneous are:
= - - = - ; and = +
10. Algorithm design proposed
.
In this study, the simulations were performed using PSCAD, and the
simulation results were obtained from the fault current signal.
The steps performed for this study were:
Finding the input to the Clarke transformation and wavelet transform. The
signal flow of PSCAD was then converted into m. files (*. M) and then
converted into mat. Files (*mat).with a sampling rate and frequency
dependent 0.5 Hz – 1 MHz .
Determining the data stream interference, where the signal was
transformed by using the Clarke transformation to convert the transient
signals into the signal’s basic current (Mode).
Transforming the mode current signals again by using DWT and WTC,
which were the generated coefficients, and then squared to be in order to
obtain the maximum signal amplitude to determine the timing of the
interruption.
Processing the ground mode and aerial mode and (WTC)2
using Bewley
Lattice diagram of the initial wave to determine the fault location
14. Simulation Model and Results
The system was connected with the sources at each end, as shown in Fig.
This system was simulated using PSCAD/EMTD. For the case study, the
simulation was modeled on a 230 kV double circuit transmission line,
which was 200 km in length. Transmission Line
Transmission data:
Z1=Z2 = 0.03574 + j 0.5776 Zo = 0.36315 +j 1.32.647
Fault Starting = 0.22 second Duration in fault = 0.15 Second
Fault resistance = 0.001 , 25, 50, 75 and 100 ohm
Fault Inception Angle = 0 , 15, 30 , 45 ,60, 90 , 120 and 150 degree
Source A and B Z1 = Z2 = Zo = 9.1859 + j 52.093 Ohm
28. The obtained result for different inception fault using DWT
and BPNN with configuration (12-24-48-4)
29. The comparison result for model BPNN and PRN based on Clarke’s
transformation with configuration (12-24-48-4)
30. The comparison SE for model BPNN and PRN based on Clarke’s
transformation
31. VE comparison for model BPNN and PRN based on
Clarke’s transformation
32. Comparison of MSE and MAE for Back Propagation
Neural Network, Pattern Recognition Network and Fit
Network Algorithm
33. This paper proposes a technique of using a combination of discrete
wavelet transform (DWT) and back-propagation neural networks (BPPN)
with and without Clarke’s transformation, in order to identify fault
classification and detection on parallel circuit transmission lines. This
technique applies Daubechies4 (Db4) as a mother wavelet. Various case
studies have been studied, including variation distance, the initial angle
and fault resistance. This study also includes comparison of the results of
training BPPN and DWT with and without Clarke’s transformation, where
the results show that using Clarke’s transformation will produce smaller
MSE and MAE, compared to without Clarke’s transformation. Among the
three structures, the Architects result was the best, which was 12-24-48-
12. Four statistical methods are utilized in the present study to determine
the accuracy of detection and classification faults, suggesting that the
Back Propagation Neural Network results in the lowest error thus it is the
best compared with Pattern Recognition Network and Fit Network.
Conclusion