UTILIZATION OF SUPPORT VECTOR MACHINE FOR FAULT DETECTION AND CLASSIFICATION IN POWER TRANSMISSION LINE SYSTEMS

By

A Thesis

Submitted to the

Faculty of Western Michigan University

in partial fulfillment of the

requirements for the

Degree of [name of degree here]

[Name of department or unit here]

Advisor: [First and Last Name, Ph.D. or Ed.D. here]

Western Michigan University

Kalamazoo, Michigan

[Month and year of graduation, no comma separating]

UTILIZATION OF SUPPORT VECTOR MACHINE FOR FAULT DETECTION AND CLASSIFICATION IN POWER TRANSMISSION LINE SYSTEMS

Western Michigan University, [year of graduation]

ABSTRACT

Electrical transmission lines are prone to faults and failures. When a fault occurs, it is impossible most of the times to fix it manually. Many methods have been adopted in the past in order to serve the purpose as fault diagnosing application. These methods mainly include; expert system, fuzzy theory, Artificial Neural Network, and Support Vector Machine (SVM). This thesis discusses the method Support Vector Machine for fault diagnosis. SVM has the edge of good generalization over other fault diagnosing applications because it is based on the principle of risk minimization principle. Due to this ability, SVM application has been widely used in fault diagnosis and classification in recent times. This thesis studies the utilization of support vector machine as tool for fault detection and classification in power transmission line systems. The aim is to identify the type of fault in the lines and to determine which segment of the line has experienced a fault. Furthermore, in this work, the current and voltage of each phase are sampled, calculated and then utilized as an optimal learning pattern. Using this method, experimental simulations will show that SVM can classify each class more accurately in comparison to previously used methods such as; Expert System, Artificial Neural Network, Petri Net, Fuzzy Theory. The results of simulation tests demonstrate the effectiveness of the proposed automatic fault diagnosis method.

ACKNOWLEDGMENT

I would like to take this chance for thanking my research facilitator, friends & family for support they provided and their belief in me as well as guidance they provided without which, I would have never been able to do this research.

DECLARATION

I, (type your full first names and surname here), would like to declare that the Project Description in the deliverable form is all my individual work without any aid and that this thesis has not been submitted for any examination at academic as well as professional level previously. I have acknowledged all quotations from the published or unpublished works of other people. It is also representing my very own views and not essentially which are associated with university.

Signature:

Date:

TABLE OF CONTENTS

ABSTRACTii

ACKNOWLEDGMENTiii

DECLARATIONiv

CHAPTER 1: BACKGROUND AND RESEARCH OBJECTIVES1

1.1 Introduction1

1.2 Research Motivation3

1.3 Research Objectives4

1.4 Research Questions4

1.5 Pertinent Literature5

1.6 Key Definitions7

1.6.1 Protection Systems7

1.6.2 SVM7

1.6.3 SVM Regression7

1.6.4 One-Class SVM7

1.6.5 Class Weights8

1.6.5 Optimization in SVM8

1.7 Ethical Considerations8

1.8 Thesis Outline9

BIBLIOGRAPHY10



CHAPTER 1: BACKGROUND AND RESEARCH OBJECTIVES

1.1 Introduction

With the rapid development of power electronics technology, fault detection and localization are the focus of research efforts in the area of transmission and distribution system. Because faults in electrical power systems cannot be avoided, enough information provided from the fault analysis is ...