review Essay

Review Report on Fault Detection an d Classification

Methodologies o n Overhead Transmission Lines

Sravan Kumar Kotha (DT18EEE086), Dr. Bhooshan Rajpathak

[email protected] , [email protected]

Department of Electrical Engineering

Visvesaraya National Institute of Technology, Nagpur.

ABSTRACT: Transmission line is the one which transmits power from generating stations

to load; hence it is the most crucial elements in the power system network . Faults on

transmission lines not only affect the equipment but also the power quality. Disturbances

(Faults) which affect the transmission line system should be detected accurate ly and

pro mptly. Fault investigation on transmission lines is imperative for economic operation of

power system components as it facilitates quicker repair, improve system availability, reduce

operating costs, and saves time and space. Various fault clas sification m ethodologies for

detection and classification of faults on transmission lines has been presented in this paper . In

this review almost all the techniques and philosophies are discussed in the literature.

Key Words: Three phase two terminal transmission line , wavelet transform, Artificial

Neural Networks (ANNs), Support Vector Machine (SVM), feature extraction, fau lt

classification accuracy .

Abbreviations: DWT: Discrete wavelet transform, ANN: Artificial Neural Networks ,

SVM: Support Vector Machine .

Introduction:

For proper functioning of power system network the protection of transmission lines

against uncovered deficit is the most basic task . Fault on transmission system is an

abnormal condition when the system quantities like voltage, current and phase angle etc

exceeds its threshold values. Faults on overhead lines can be identified by examining the

phase voltages and phase currents as open circuit (series) faults and the ot her one is short

circuit (shunt) faults . Short circuit faults again classi fied into LG, LL, LLG, and LLL faults .

Here L indicates each phase and G indicates ground. Among all the faults m ost severe fault

on transmission line is LLL fault or 3 -phase .

For maintaining the reliability of power system continuous operation of transmis sion line

system is having prime importance . Disturbance (Fault) which affects the transmission line

should be detected accurately and fast. Accurate fault classification facilitates quick repair,

improves avai lability of system , reduce maintenance cost during operation , and saves time.

Various approaches are used for fault type identification and classification . Every

methodology has its merits and flaws. So it is difficult to select a n accurate fault classificatio n

for users. Hence, an effort is made in this paper t o make a review of all the efficacious fault

classification method s. In this review, fault classification methods have been related based on

their techniques and simulation tools used. The power system model considered is simulated

by using di fferent tools like MATLAB, EMTP, EMTDC/PSCAD.

Survey on Fault Classification Techniques :

In this review paper popular and hybrid techniques have been discussed. In addition to

this newly proposed methodologies are also presented for the ease to approach.

Popular methodologies

Popular methodologies used for detection and classification of fault types. They are

1. Wavelet -Transform (WT) approach

2. Artificial Neural Network approach (ANN)

3. Support Vector Machine Technique (SVM) .

The above techniques are explained below.

1. Wavelet -Transform (WT) Approach

In earlier days Fourier transforms were used to extract features of the signals in frequency

domain . There may be a possibility of losing time information by using Fourier transform.

This may be overcome by using Mul ti resolution analysis (MRA ) in Wavelet transforms

which can decompose the signal into different frequency bands . It allows the decomposition

of signal into various resolution levels. This is done by implementing a discrete wavelet

tran sform (DWT) using a bank of HP and LP filters .

To implement wavelet MRA, the selection of Mother Wavelet [1] is essential. An attempt

was made by using Db4 mother wavelet and the use of discrete wavelet transform (DWT) for

classification of faults on transmission lines . In th is method fourth level detailed coefficients

of the three phase fault current signals have been used.

An approach has been made by utilising discrete wavelet transform (DWT) for classifying

faults on ove rhead lines [2 ]. In this method the detailed an d approximation coefficients of

three phase currents and voltages are utilises to identify the occurrence of faul t on lines and

the type can also be identified .

2. Artificial Neural Network Approach

Earlier fault classification was done by phase selector, but was unable to classify the correct

fault type under different system operating conditio ns like fault resistances. Artificial Neural

networks (ANNs) can map abnormal patterns which intern provides potential solution to the

long -standing problems of accurat e fault classification .

Back propag ation (BP) neural system is utilised in ANN for fault classification [ 3]. A feed

forward network with back propagation algorithm by the use of radial basis function [4] has

been developed for cla ssification of faults on d ouble c ircuit transmission lines.

3. Support Vector Machine (SVM ) Approach

Support Vector Machine optimally se parates data into two groups by constructing an N –

dimensional hyper plane. SVM model using a sigmoid kernel functions which are similar to

a two layer perceptron neural -networks. From detailed coefficients of current signals obtained

from wavelet transform some number of pre -faulted samples and some number of post –

faulted samples are taken from each signal sample for training of SVM. Pre -faul ted samples

are given with zero label and post -faulted samples are given with some integer valued label

for ex. 1 for AG fault, 2 for BG fault etc. After extracting the useful features from detailed

coefficients of three current signals are used as input to the SVM .

Hybrid Techniques

Hybrid techniques utilises the comb ination of above three methodologie s. Concoction has

been done to mitigate the deficiency in one approach during its application.

In [5], the author classifies the faults on transmission lines in three stages. From sending end

of transmission line t he current and voltage signals were obtained in first stage and these are

processed to get detailed coefficients by discrete wavelet tran sform (DWT). In second and

third stages multi class support vector machine (MCSVM) is employed for classification and

location fault.

The feature extraction of faulty phase current and voltage signals is done by a hybrid

technique [6] by utilising discrete wavelet transform (DWT). These are used to train the SVM

classifier to make a decision of fault or no fault. Ground detection is done by using proposed

ground index.

Global positioning system (GPS) has been in troduced [7] for fault identification and

categorisation on overhead lines. Wavelet transform (WT) is used to obtain detailed

coefficients which are utilized to get fault indices. Threshold value is set based on

enormous test signals. If fault indices are compared with this threshold values for fa ult

classification. ANN is used to locate the faults.

A new approach has been discussed in [8] ; The features of faulty phase currents were

decomposed to first level by using discrete wavelet transform and these are given as inputs

to train and test support vector machine and extreme learning machine ELM.

The presence of non -linear loads has been considered while detecting and classifying faults

on transmission line [9] by using SVM and ANN techniques. T o estimate the harmonic

components Kalman filter is use d to process the post fault voltage signals, which can be

served as input vectors for SVM and ANN.

To find the distance of fault from generating end is proposed in [10]. The wavelet

technique and neural network techniques were used. This method produces ve ry

imperative fault distance results with a relative error around 3.2%.

Conclusion:

Availability of fault detection and classification methodologies have been presented in

thi s review report which are helpful for selecting the basic technologies and

amalgamation of different technologies used for protectio n of overhead lines. This report

explains the different techniques based on the technique and simulation tools used i.e.

popular and combination of b asic techniques. One can extend the work by utilisi ng the

methods discussed in this article.

Acknowledgement: I convey my deep sense of gratitude to VNIT Nagpur for providing me

the opportunity to do research.

References:

1) P. Rajaraman, N.A. Sundaravaradan, Rounak Meyur, M. Jaya Bharatha Reddy, and

D.K. Mohanta Fault Clasification in Transmission Lines using Wavelet Multi

2) V. Ashok, K. G. V. S. Bangarraju & V. V. N. Murthy Identification and Classification

of Transmission Line Faults Using Wavelet Analysis ISSN ( PRINT) : 2320 8945,

Volume -1, Issue -1, 2013 .

3) Eisa Bashier M. Tayeb Orner AI Aziz AlRhirn Transmission Line Faults Detection,

Classification and Location using Artificial Neural Network 978 -1-4673 -6008 –

4) Mr. N.Saravanan, Mr. A.Rathinam A Comparative Study on ANN Based Fault

Location and Classification Technique For Double Circuit Transmission Line, 2012

Fourth International Conference on Computational Intelligence and Communication

Networks .

5) Sami Ekici, Support Vector Machines for classifi cation and locating faults on

transmission lines Applied Soft Computing 12 (2012) 1650 1658 .

6) Manohar Singh, Bijaya Ketan Panigrahi, R.P. Maheshwari Transmission line fault

detection and classification; PROCEEDINGS OF ICETECT 2011, 978 -1-4244 -7926 –

7) Abdul Gafoor Shaik, Ramana Rao V. Pulipaka A new wavelet based fault detection,

classification and location in transmission lines, Electrical Power and Energy Systems

64 (2015) 35 40, .

8) V. Malathi, N.S. Marimuthu, S. Baskar Intelligent approaches using support vector

machine and extreme learning machine for transmission line protection

Neurocomputing 73 (2010) 2160 2167, 0925 -2312 .

9) Ebha Koley, Sunil K. Shukla, Subhojit Ghosh, Dusmanta K. Mohanta Prot ection

scheme for power transmission lines based on SVM and ANN considering the

presence of non -linear loads IET Generation, Transmission & Distribution, IET

Gener. Transm. Distrib. , 2017, Vol. 11 Iss. 9, pp. 2333 -2341 .

10) Majid Jamil, Md.Abul Kalam, A.Q.An sari, M.Rizwan Wavelet ? FFNN Based Fault

Location Estimation of a Transmission Line, Electrical Engineering Research Vol. 1