ACKNOWLEDGEMENT

A journey is easier when you travel tougher. Interdependence is certainly more valuable than independence. The real spirit of achieving a goal is through the way of excellence and perpetual discipline. I would have never succeeded in completing my task without the cooperation, encouragement and help provided to me by various personalities. First of all, I render my gratitude to the almighty who bestowed self-confidence, ability and strength in me to complete this work. Without his grace this would have never been a reality. With deep sense of gratitude I express my sincere thanks to my esteemed and worthy Supervisor Dr. CHIRANJIB KOLEY, Associate Professor, Electrical Engineering Department for his valuable guidance in carrying out this work under his effective supervision, encouragement, enlightenment and cooperation. I am grateful to Dr. N.K.ROY, Head of the Department Electrical Engineering for his constant encouragement that was of great importance in the completion of this thesis. I am also thankful to all the staff members of the department for their full cooperation and help. My greatest thanks to all who wished me success especially my parents and friends whose support and care made me stay on earth.

Place: Durgapur Date: (Rajashree Dhua) Roll No: 10/EE/405

NATIONAL INSTITUTE OF TECHNOLOGY, DURGAPUR DECLARATION

I hereby declare that this submission is my own work and that, to the best of my knowledge and belief, it contains no material previously published or written by another person nor material which has been accepted for the award of any other degree or diploma of the university or other institute of higher learning, except where due acknowledgement has been made in the text.

Place: N.I.T. Durgapur Date:

Signature Name Roll No.

: : Rajashree Dhua : 10/EE/405

ii

NATIONAL INSTITUTE OF TECHNOLOGY, DURGAPUR CERTIFICATE

This is to certify that Rajashree Dhua (Roll Number 10/EE/405), undergoing Master of Technology in Electrical Engineering, with specialization in Electrical Systems has carried out the dissertation titled Simulation and Identification of Transmission Line Faults and prepared the report under my guidance and supervision. The dissertation is submitted as a partial fulfilment of the requirement for the award of Master of Technology in Electrical Engineering with specialization in Electrical Systems from National Institute of Technology, Durgapur. To the best of my knowledge, the materials in this report have not been submitted earlier as a part of any other academic programme.

Dr. Chiranjib Koley Professor and Supervisor Department of Electrical Engineering National Institute of Technology, Durgapur

Dr. N.K. Roy Professor and Head Department of Electrical Engineering National Institute of Technology, Durgapur

iii

NATIONAL INSTITUTE OF TECHNOLOGY, DURGAPUR CERTIFICATE OF APPROVAL

The foregoing thesis entitled Simulation and Identification of Transmission Line Faults is hereby approved as a creditable study of an Engineering project carried out and presented in a manner satisfactory to warrant its acceptance as prerequisite to the degree for which it has been submitted. It is understood that by this approval, the undersigned do not necessarily endorse any conclusion or opinion therein, but approved the thesis for the purpose for which it is submitted.

Examiner

... Examiner

Examiner

iv

ABSTRACT Though transmission lines are designed to ensure a reliable supply of energy with the highest possible continuity, but about 85-87% of faults in power system occur in transmission lines. Faults can occur due to external causes or internal failures in the power system. Identification of type of faults as well as location of faults is extremely necessary to reduce the outage time and maintenance works. Fault identification is a difficult task because practical experimental verification is difficult; also there is no standard method for identification. Switching phenomenon occurring in transmission lines often produces similar types of transients as that of faults, making the identification task even more difficult. The different types of faults occurring in transmission lines can be categorized as unsymmetrical faults such as ground fault (LG), line to line fault (LL) and double line to ground fault (LLG) and symmetrical faults such as three phase fault (LLL) and three phase to ground fault (LLLG). Apart from the symmetrical and unsymmetrical faults, arc faults also occur in the power system, which may be a static arc fault or a dynamic arc fault. Switching transients occurring in transmission lines due to various reasons also produce similar kind of transient waveforms. The major reasons for the occurrence of faults are external environmental conditions like storm, sudden fall of a tree branch and also internal causes like insulation failure, breakdown of insulator, faulty tripping of a circuit breaker. In practical scenario fault can occur at any location, any time (any inception angle) and the fault resistance can vary from as low as few ohms to few hundreds of ohms, which influence the transient characteristics of the voltage waveforms irrespective of the type of fault. This makes the fault type identification, a difficult task. Therefore, the main objective is to identify the

type and possible location of fault, for which an accurate digital model of a test system becomes necessary, as practical experimentation is difficult. In the work, a transmission line has been modelled, using distributed parameter, as it has been found to be more accurate for high frequency transient study. In order to make the transmission line model closer to the actual transmission line, the parameters of the transmission line such as R, L and C has been considered frequency dependent, instead of being constant as in the case of the constant parameter model. The digital model of the proposed system has been implemented with the help of Electro Magnetic Transients Program (EMTP), which is freely available and widely used software for transient studies. The study of the transient characteristics of voltage waveforms at different fault conditions, reveals that, the transient waveform for different types of fault (with same fault parameter) are different, but variation of the fault parameters like fault resistance, location, fault inception angle influence the transient characteristics in a similar manner, making the fault identification a difficult task. As the recorded voltage waveforms for different fault condition are nonstationary in nature, i.e. the harmonic content changes with time, the Short Term Fourier Transform (STFT) has been performed in order to study the variation of the transient behaviour closely in time-frequency domain. Through time-frequency domain studies, it has been observed that arc faults and switching transients can be easily identified from the other faults because of their distinctly different frequencies and amplitude. Though, symmetrical and unsymmetrical fault identification remains a difficult task, because these form a complex relationship with interdependence and overlapping in terms of frequencies. Finally through statistical analysis, a threshold value has been

vi

identified for different type of faults and the proposed work can be further extended with the implementation of a suitable classifier through training and testing methods. The work takes into account most of the common disturbances that occur in transmission lines and develops a test system for fault simulation based on the frequency dependent transmission line model. The work also proposes a method for identification of fault type and location using STFT.

vii

TABLE OF CONTENTS
ACKNOWLEDGEMENT DECLARATION CERTIFICATE CERTIFICATE OF APPROVAL ABSTRACT TABLE OF CONTENTS LIST OF FIGURES LIST OF TABLES i ii iii iv v viii xi xv PAGE NO

CHAPTER 1: INTRODUCTION
1.1 Previous work on identification of various faults 1.2 Objective 1.3 Work summary 1.4 Thesis organization 1 2 2 3

CHAPTER 2: MODELING OF TEST SYSTEM

2.1 Overview of digital fault simulators 2.2 Various models for fault simulation 2.3 Lumped parameter model for fault simulation 2.4 Distributed parameter model for fault simulation 2.5 Test system for fault simulation 2.5.1 Line parameters 4 4 5 6 9 9

CHAPTER 3: SIMULATION OF FAULTS

3.1 Comparison between transients in a distributed parameter model and a lumped parameter (PI) model
viii

10

3.2 Types of faults 3.3 Causes of faults 3.4 Arc faults 3.5 Switching over voltages 3.6 Different types of fault simulation 3.6.1 Variation of parameters for fault simulation 3.6.2 Arc fault simulation

10 11 11 12 12 12 13

CHAPTER 4: TIME DOMAIN ANALYSIS

4.1 Simulation parameters 14

4.2 Simulation results for different unsymmetrical and symmetrical faults with variation of fault resistance, fault location and fault inception angle 4.2.1 Observations 4.3 Dynamic arc fault simulation 4.4 Switching over voltage simulation 14 22 22 24

CHAPTER 5: FREQUENCY DOMAIN ANALYSIS

5.1 Power spectral density (PSD estimate) 5.2 Transfer function estimate 26 28

CHAPTER 6: TIME FREQUENCY DOMAIN ANALYSIS

6.1 Short time Fourier transform (STFT) 6.1.1 Continuous-time STFT 6.1.2 Discrete-time STFT 6.2 Spectrogram analysis of various fault signals 6.2.1 Spectrograms for line to ground(AG) fault 6.2.2 Spectrograms for line to line (AB) fault
ix

30 30 31 31 32 36

6.3 Dynamic arc fault spectrograms 6.4 Switching transients spectrogram

41 42

CHAPTER 7: FEATURE EXTRACTION FOR FAULT IDENTIFICATION

7.1 Feature analysis from spectrograms 7.1.1 Variation of frequency of first peak with fault location 7.1.2 Variation of amplitude of first peak with window number 7.1.3 Variation of amplitude of first peak with window number for different fault resistance 7.1.4 Variation of slope with fault resistance 7.1.5 Observations 45 46 47 43 43 44

7.2 Estimation of different fault parameters and feature extraction from an unknown time domain signal 7.2.1 Evaluation of fault inception angle by Discrete Wavelet Analysis 7.2.1.1 Wavelet transform and Discrete Wavelet transform 7.2.2 Estimation of fault resistance and fault location 47 47 47 49

CHAPTER 8: RESULT ANALYSIS

8.1 Variation of Amplitude and Frequency for different type of faults 50

8.2 Statistical analysis and box plot for different symmetrical and unsymmetrical faults

CHAPTER 9: CONCLUSION AND SCOPE OF FUTURE WORK

9.1 Conclusion 9.2 Scope of future work 55 55

REFERENCES
x

56

LIST OF FIGURES
Fig 2.1 Lumped parameter PI model (single section) Fig 2.2: Unfaulted Long Transmission Line Fig 2.3: Transmission line fed from one end Fig 3.1 Voltage transients generated due to a ground fault (AG fault) in a distributed parameter line and a lumped parameter (pi) modelled 100 km line Fig 3.2 Pie chart showing percentage of occurrence of faults

Fig 4.1:Voltage waveforms for line to ground fault (AG fault) with variation of fault resistances(1 , 10 , 50 and 100 ) for fault inception angle 20 and fault location 20 km Fig 4.2: Voltage waveforms for line to ground fault for fault resistance 1 and fault inception angle 20 with variation in fault location(20km,40 km,60 km,80 km and 100 km)

Fig 4.3: Voltage waveforms for a line to ground fault at 20 km length of the transmission line and fault resistance (Rf)=1with variation in fault inception angle(20,90 and 135)

Fig 4.4: Voltage waveforms for a line to line fault (AB fault) at 20 km length and fault inception angle (FIA) =20 with variation of fault resistance (10, 50, 100 and 200 ) Fig 4.5:Voltage waveforms for a line to line fault for fault resistance 10 and fault inception angle 20 with variation in fault location (20 km,40 km,60 km,80 km and 100 km) Fig 4.6: Voltage waveforms for a line to line fault at 20 km and fault resistance10 with variation in fault inception angle as(20,90,180 and 225)

xi

Fig 4.7: Voltage waveforms for a double line to ground (ABG fault) at 20 km length and fault inception angle (FIA) =20 with variation of fault resistance (10 , 50, 100 and 200) Fig 4.8:Voltage waveforms for a double line to ground fault for fault resistance 50 and fault inception angle 20 with variation in fault location (20 km,40 km,60 km,80 km and 100km)

Fig 4.9: Voltage waveforms for a double line to ground fault at 20 km and fault resistance 50 with variation in fault inception angle as(20,90,180 and 225)

Fig 4.10: Voltage waveforms for a three phase fault (ABC fault) at 20 km length and fault inception angle (FIA) =20 with variation of fault resistance (10 , 50, 100 and 200) Fig 4.11:Voltage waveforms for a three phase fault for fault resistance 100 and fault inception angle 20 with variation in fault location (20 km,40 km,60 km,80 km and 100km) Fig 4.12: Voltage waveforms for a three phase fault at 20 km and fault resistance100 with variation in fault inception angle as(20,90,180 and 225)

Fig 4.13: Voltage waveforms for a three phase to ground fault (ABCG fault) at 20 km length and fault inception angle (FIA)=20 with variation of fault resistance (10 ,50,100 and 200) Fig 4.14:Voltage waveforms for a three phase to ground fault for fault resistance 100 and fault inception angle 20 with variation in fault location (20 km,40 km,60 km,80 km and 100km)

Fig 4.15: Voltage waveforms for a three phase to ground fault at 20 km and fault resistance100 with variation in fault inception angle as(20,90,135 and 225)

xii

Fig 4.16: Line to ground voltage of phase A at the load end showing the dynamic arc fault characteristics for a line to ground fault (AG fault) at 50 km (Tdynamic=3 ms,Lstatic=0.1m)

Fig 4.17: Line to ground voltage of phase A at the load end showing the dynamic arc fault characteristics for a line to ground fault (AG fault) at 50

km(Tdynamic=0.25ms,Lstatic=0.2m)

Fig 4.18: Line to ground voltage of phase A at the load end showing the dynamic arc fault characteristics for a line to ground fault (AG fault) at 50 km(Tdynamic=0.625 ms,Lstatic=3.4m)

Fig 4.19: Network for simulation of switching over voltages due to 350 MW load shed Fig 4.20: Switching over voltages of phase A, B and C due to sudden load rejection Fig 5.1 Power spectral density estimate of AG fault voltage signal for different fault location of 25 km, 50 km and 75 km Fig 5.2 Power spectral density estimate of AB fault voltage signal for different fault location of 25 km, 50 km and 75 km Fig 5.3 Transfer function estimate of AG fault Fig 5.4 Transfer function estimate of AB fault Fig 6.1 STFT time-frequency representation Fig 6.2 Spectrogram for AG fault at 25 km, Rf =1 ohm and inception angle=90 Fig 6.3 Spectrogram for AG fault at 50 km, Rf =1 ohm and inception angle=90 Fig 6.4 Spectrogram for AG fault at 75 km, Rf =1 ohm and inception angle=90 Fig 6.5 Spectrogram for AG fault at 50 km, Rf =10 ohms and inception angle=90 Fig 6.6 Spectrogram for AG fault at 50 km, Rf =50 ohms and inception angle=90
xiii

Fig 6.7 Spectrogram for AG fault at 50 km, Rf =1 ohm and inception angle=20 Fig 6.8 Spectrogram for AG fault at 50 km, Rf =1 ohm and inception angle=135 Fig 6.9 Spectrogram for AB fault at 25 km, Rf =10 ohms and inception angle=60 Fig 6.10 Spectrogram for AB fault at 50 km, Rf =10 ohms and inception angle=60 Fig 6.11 Spectrogram for AB fault at 75 km, Rf =10 ohms and inception angle=60 Fig 6.12 Spectrogram for AB fault at 50 km, Rf =50 ohms and inception angle=90 Fig 6.13 Spectrogram for AB fault at 50 km, Rf =100 ohms and inception angle=90 Fig 6.14 Spectrogram for AB fault at 50 km, Rf =200 ohms and inception angle=90 Fig 6.15 Spectrogram for AB fault at 50 km, Rf =10 ohms and inception angle=20 Fig 6.16 Spectrogram for AB fault at 50 km, Rf =10 ohms and inception angle=135 Fig 6.17 Spectrogram for AB fault at 50 km, Rf =10 ohms and inception angle=225 Fig 6.18 Spectrogram for dynamic arc fault (AG fault) at 50 km (Tdynamic=3 ms, Lstatic=0.1m) Fig 6.19 Spectrogram for dynamic arc fault (AG fault) at 50 km (Tdynamic=0.25 ms,Lstatic=0.2m) Fig 6.20 Spectrogram for dynamic arc fault (AG fault) at 50 km (Tdyanmic=0.625 ms, Lstatic=3.4m) Fig 6.21 Spectrogram of switching transient due to load rejection at 10 ms Fig 7.1 Variation of frequency of first peak with fault location: RF = 10 for AG, ABG, ABCG and AB faults and RF=20 for ABC fault Fig 7.2 Variation of frequency of first peak with fault location: Rf= 50 for all faults Fig 7.3 Variation of amplitude of first peak with window number: Fault location=50 km, Rf=10 and FIA=90 for AG, ABG, ABCG, AB fault and fault location=50 km, Rf=10 and FIA=60 for ABC fault

xiv

Fig 7.4 Variation of amplitude of first peak with window number: Fault location=50 km, Rf=50 and FIA=90 for AG, ABG, ABCG, AB fault and fault location=50 km, Rf=50 and FIA=60 for ABC fault Fig 7.5 Variation of amplitude of first peak with window number for different fault resistance: ABC fault at fault location=50 km, fault inception angle=60 Fig 7.6 Slope calculation ABC fault at fault location=50 km, fault inception angle=60 at different fault resistances Fig 7.7: Variation of slope with fault resistance Fig 7.8 Analysis of a signal using wavelet transform Fig 8.1 Variation of amplitude with frequency for switching transients, arc faults and various symmetrical and unsymmetrical faults Fig 8.2 Variation of amplitude with frequency for symmetrical and unsymmetrical faults Fig 8.3 Variation of amplitude with frequency for all faults involving ground Fig 8.4 Variation of amplitude with frequency for faults not involving ground Fig 8.5 Box plot for different faults

LIST OF TABLES
TABLE 8.2.1: Statistical data for faults involving ground TABLE 8.2.2: Statistical data for faults not involving ground

xv