Title:Gradient Ascent Optimization for Fault Detection in Electrical Power Systems based on Wavelet Transformation
VOLUME: 15 ISSUE: 3
Author(s):Iyappan Murugesan* and Karpagam Sathish
Affiliation:Department of Electrical and Electronics Engineering, M.Kumarasamy College of Engineering, Thalavapalayam, Karur, Tamilnadu 639113, Department of Electrical and Electronics Engineering, Hindusthan College of Engineering and Technology, Coimbatore, Tamilnadu 641032
Keywords:Deep neural learning, electrical power system, fault detection optimization, feature extraction, gradient ascent, normalization.
Abstract:
Background: This paper presents electrical power system comprises many complex and interrelating
elements that are susceptible to the disturbance or electrical fault. The faults in electrical
power system transmission line (TL) are detected and classified. But, the existing techniques like Artificial
Neural Network (ANN) failed to improve the Fault Detection (FD) performance during transmission
and distribution. In order to reduce the Power Loss Rate (PLR), Daubechies Wavelet Transform
based Gradient Ascent Deep Neural Learning (DWT-GADNL) Technique is introduced for FDin electrical
power sub-station. DWT-GADNL Technique comprises three step, normalization, feature extraction
and FD through optimization. Initially sample power TL signal is taken. After that in first step,
min-max normalization process is carried out to estimate the various rated values of transmission lines.
Then in second step, Daubechies Wavelet Transform (DWT) is employed for decomposition of normalized
TL signal to different components for feature extraction with higher accuracy. Finally in third step,
Gradient Ascent Deep Neural Learning is an optimization process for detecting the local maximum (i.e.,
fault) from the extracted values with help of error function and weight value. When maximum error
with low weight value is identified, the fault is detected with lesser time consumption. DWT-GADNL
Technique is measured with PLR, Feature Extraction Accuracy (FEA), and Fault Detection Time
(FDT). The simulation result shows that DWT-GADNL Technique is able to improve the performance
of FEA and reduces FDT and PLR during the transmission and distribution when compared to state-ofthe-
art works.
Materials and Methods: An electric power system incorporates production, broadcast and distribution
of electric energy. To send the electric power to massive load centers, transmission lines are exploited.
The fast growth of electric power systems results in huge number of lines in operation and total length.
TL are susceptible to faults in case of lightning, short circuits, mis-operation, human errors, overload, etc.
Faults resulted in tiny to long power outages for customers. To protect the reliable power system operations,
Fault identification, isolation and localization are imperative. The voltage lessened to minimal value,
when fault occurs on TL. FD is an essential problem in power system engineering to minimize the PLR.
DWT-GADNL Technique is introduced for FD in TL during transmission and distribution.
Results: Power Loss due to the fault occurrence during the transmission and distribution is a common
problem in electrical power system. To lessen the PLR, the fault is detected in earlier stage. From the
sample transmission line, the features are extracted and the values are calculated. When the observed
value is lesser than the actual value, the fault is detected through performing the gradient ascent optimization
process in transmission line. In this optimization process, the local maxima are identified to reduce
the PLR. At different time instances, PLR gets changed. At instance 3, the PLR of proposed DWTGADNL
framework is 12% where the PLR of Fuzzy Logic Based Algorithm and Fault Diagnosis
Framework are 27% and 19% respectively. Through comparing all the ten instances, PLR is reduced in
GWMD-DE technique by 59% and 40% compared to existing respectively.
Conclusion: DWT-GADNL Technique is introduced for FD during transmission and distribution with
minimal PLR. Sample power TL signal is taken and min-max normalization process performs the various
rated values estimation of transmission lines. DWT decomposes normalized TL signal to different
components for feature extraction with higher accuracy. Gradient Ascent Deep Neural Learning detects
the local maximum from extracted values with help of error function and weight value. When maximum
error with low weight value is identified, the fault is detected with lesser time consumption. The performance
of DWT-GADNL technique is tested with the metrics such as PLR, FEA and FDT. With the
simulations conducted for all techniques, the proposed DWT-GADNL technique presents better performance
on FD during transmission and distribution as evaluated to state-of-the-art works. From simulations
results, the DWT-GADNL technique lessens PLR by 50% and enhances FEA by 9% than the existing
methods.