Fault Location on Transmission Lines Based on

Travelling Waves
F. V. Lopes, Student Member, IEEE, D. Fernandes Jr., Member, IEEE, W. L. A. Neves, Member, IEEE

AbstractAutomatic and accurate fault location methods for

transmission lines may reduce the search time technicians would
take to find out where the fault is, leading to a quick recovery of
the system. Usually, transient detection techniques use more than
one sample of voltage or current to make possible the required
transient analysis. This paper presents a very simple method for
fault location on transmission lines based on travelling waves
using Parks Transformation to perform the transient detection.
An advantage of the proposed method is that only one voltage
sample at each phase is used and transients in all three phases are
monitored simultaneously by the analysis of only one signal: the
direct axis voltage. The method is implemented in the Alternative
Transients Program (ATP) using the MODELS language and
simulations were carried out for a 230 kV transmission line with a
digital fault recorder (DFR) at each terminal to evaluate the
applicability of the proposed technique. The results show the
effectiveness of the algorithm which can be used as an additional
routine for digital protective relays.
Keywords: Fault location, Parks Transformation, transient
detection, transmission lines.

I. INTRODUCTION

HE exact fault location on transmission lines is a problem

that has been studied for decades. The theory of travelling
waves has been recently used for this purpose. Such methods
are based on the transient detection at specific points of the
electrical system in question and are classified according to the
number of monitored points. Several methods for transient
detection have been proposed in order to enable an accurate
and reliable fault location for transmission lines [1].
Methods based on travelling waves theory utilize voltage
and/or current data captured by DFRs (Digital Fault
Recorders) installed in monitored line terminals. According to
[2], for double ended fault location methods, it is necessary to
use GPS (Global Positioning System) in order to synchronize
oscillographic records at both monitored ends. So, digital
signal processing techniques are applied to synchronized data
from DFRs, making possible the detection of initial transient

This work was supported by Brazilian Coordination for the Improvement of

Higher Education Personnel (CAPES).
F. V. Lopes is M. Sc. candidate at the Federal University of Campina Grande
(UFCG), Av. Aprgio Veloso, 882, Bodocong, 58.429-140, Campina
Grande-PB, Brazil (e-mail: [email protected]).
D. Fernandes Jr. and W. L. A. Neves are with Department of Electrical
Engineering
of
UFCG.
(e-mail:
[email protected],
[email protected]).
Paper submitted to the International Conference on Power Systems
Transients (IPST2011) in Delft, the Netherlands June 14-17, 2011

instants at both transmission line ends. Then, fault location

may be determined using expressions proposed in [3] in which
the input data are the propagation velocity of travelling waves,
the line length and the initial transient instants in monitored
terminals.
In [4], the Discrete Wavelet Transform (DWT) and the
Redundant Discrete Wavelet Transform (RDWT) are
presented as efficient techniques for transient detection. In [5]
is proposed a fault location method based on the analysis of
correlation coefficients which are calculated from the voltage
oscillographic records in both line ends.
The accuracy of travelling wave methods is a function of
the sampling frequency of DFRs, i.e., it depends on the
hardware used for data acquisition. In fact, it is rather refer to
reliability than accuracy and define a maximum expected error
as a function of the sampling frequency.
It is important to know that the transient detection is a
critical step for fault location procedure and, thus, the
reliability of the fault point estimation by any algorithm
depends directly on the reliability of the method used to
estimate the initial transient instants. In this paper, a very
simple method for transient detection is proposed.
Conventional methods usually need a window with more than
one sample of current or voltage waveforms to perform the
required transient analysis. Here, the window has only one
sample, i.e., no previous voltage samples are needed. All three
phases are monitored by the direct axis voltage signal Vd
obtained through Parks transformation (Tdq0) and the method
automatically determines the fault location point immediately
after the fault occurrence.
II. DOUBLE ENDED FAULT LOCATION METHODS
The main difference among travelling wave methods is the
number of monitored terminals and the technique to detect the
transients. In [3] are presented expressions for determination
of fault location considering one and two transmission line
monitored ends. In [6] is proposed a three-terminal method
that avoids the utilization of the propagation velocity of the
travelling waves in the fault location procedure. In order to
eliminate the propagation velocity variable, three terminals of
two neighbor transmission lines are monitored and a unique
value of propagation velocity is utilized for both lines. The
problem of such method is that the waves present different
propagation velocities depending on the ground resistivity and
line parameters and, consequently, the utilization of the same
propagation velocity for both lines can be considered as a
source of error. Other works, as [7] and [8], present single

ended fault location methods. Such method use incident and

reflected waves to perform the fault location.
Basically, the key idea of all mentioned methods is the
determination of initial transient instants in monitored
terminals. Here, the fault location method is based on the
analysis of voltage waveforms recorded by DFRs in two
transmission line terminals and, so, it is considered to be a
double ended method. In this way, only incident travelling
waves are used to perform the fault location, avoiding the
utilization of reflected waves at the fault point and,
consequently, making the method more reliable. A GPS must
be used in order to synchronize the DFRs measurements to
bring voltage samples to the same timebase and to ensure that
the initial transient instants will be correctly obtained.
Regarding the GPS usage, it is important to know that, in
practice, it can present synchronization errors in the order of
1s. In spite of such limitation, it is possible to use an
independent clock on the fault location device that initiates the
counter at the moment of the first transient detection and stop
the counter at the second transient detection at the opposite
bus. In this way, synchronization errors are minimized. To
understand the principles of travelling wave methods, consider
Fig. 1 in which is shown DFRs placed at Buses 1 and 2. A time
space diagram (Lattice diagram) is shown in the middle of the
figure. A scheme of the DFRs and GPS application in the
proposed double ended method is also presented in Fig. 1.

it is necessary to detect the first incident waves at buses 1 and

2 to compute the time instants t11 and t21. Secondly, the fault
point location is obtained from an expression proposed in [3]:
l + (t11 t 21 ) v ,
(1)
2
where v is the propagation velocity of travelling waves and l is
the line length. The procedure to perform the transient
detection is explained next.
d =

III. TRANSIENT DETECTION USING PARKS TRANSFORMATION

The proposed method for transient detection is based on the
work presented by R. H. Park in the United States in 1929 [9].
Parks transformation (Tdq0) has been used in the whole area
of electrical engineering, especially on researches regarding to
salient pole synchronous machines where variable inductances
in a static reference frame are seen as constant inductances
when the reference frame rotates at synchronous speed. Here,
the rotating reference system is used to remove power
frequency signals from voltage waveforms allowing transient
time detection. An analogy between the Tdq0s application in
electrical machines research and in transient detection is
shown in Fig. 2.

(a)

Fig. 1. Double ended fault location method scheme.

Where:
t11 is the initial transient instant at Bus 1;
t21 is the initial transient instant at Bus 2;
t12 is the arrival instant of the reflected wave at Bus 1;
t22, t23 are the arrival instants of reflected waves at Bus 2;
t22r is the arrival instant of the refracted wave at Bus 1;
l is the line length;
d is the distance to the fault point.
The t21r instant was left off in Fig. 1 to simplify the diagram.
It is important to know that fault location algorithms consist on
two main steps the transient detection and the fault point
location. For the first step, considering double ended methods,

(b)
Fig. 2. Parks Transformation (Tdq0) application: (a) Electrical machine
researches; (b) Transient detection and fault location.

For power system operating at normal conditions, zero

frequency signals are calculated and, if transients occur,
oscillatory signals are obtained. Such signals are called direct
and quadrature axes components which will be represented
from now on as Vd and Vq. Both components may be used for
transient detection but, here, only Vd will be considered.
For high impedance fault cases, Vd coefficients presents
high attenuation. Thus, to increase the sensitivity of the

proposed algorithm, difference coefficients (cdif) are calculated

using Taylors approximation:
c dif (i ) =

V d (i ) V d (i 1) .
t

(2)

Where:
Vd is the direct axis component;
i is the sample number;
t is the time step.
In the literature, difference quantities are widely used in
protective relaying algorithms [10]. In this work, [cdif]2 will be
used to detect de initial transient instants in monitored
terminals. [cdif]2 makes the transient detection more robust
because coefficient related to transient signals are amplified
and coefficients related to normal conditions of the system are
kept with low magnitude. An example of transient detection
using Tdq0 is shown in Fig. 3.

reference frame, (c) calculation of an orthogonal voltage

phasor in a rotating reference frame, (d) determination of the
initial transient instants and (e) fault location.
A. Data Acquisition Using DFRs in Both Line Ends
Here, only voltage signals are used. An important
characteristic of data acquisition is related to the sampling
frequency utilized by DFRs. The low sampling rates of
commercially available DFRs (usually in between 15 kHz and
20 kHz) is a limitation for fault location methods based on
travelling waves. A 50 s time step was used in ATP
simulations to simulate DFRs with 20 kHz sampling frequency
(333 samples/cycle). Keeping the number of samples/cycle the
same, smaller time steps would be even more appropriate to
reduce errors produced by numerical integration methods.
Anyway, even with this limitation, algorithms based on
travelling waves are more reliable than those based on
fundamental frequency components.

x 10

B. Orthogonal Voltage Phasors (Static Reference Frame)

The Tdq0s input variables are orthogonal phasors obtained
through Clarkes transformation which calculates the aerial
modes (V, V) and the ground mode (V0):

(a)
Three-phase
Voltage Signal (V)

2
PhaseA
1

PhaseB

PhaseC

-1

(b)
Direct Axis Componente - V d (p.u.)

-2
0.8

0.9

1.1

1.2
1.3
1.4
Sample Number

1.5

1.6

1.7

1.8
4

x 10

0.5

-0.5
0.8

0.9

1.1

1.2
1.3
1.4
Sample Number

1.5

1.2
1.3
1.4
Sample Number

1.5

1.6

1.7

1.8
4

x 10

(c)
Difference Coefficients
c dif (p.u.)

x 10
3
2
1
0
-1

(d)
Square of the Difference Coefficients
[c dif ]2

0.8

0.9

1.1

1.6

1.7

1.8

10

0
0.9

1.1

1.2
1.3
1.4
Sample Number

1
2

2
1

3
0

1
1
3

2 V A
.
1 V B
2

3 V C
2
1

2
2

(3)

Here, only aerial modes are used because their propagation

velocity is higher than those of ground mode waves. V and V
are used as Tdq0s inputs variables in order to obtain the
orthogonal voltage phasors in a rotating reference frame. Note
that Clarkes transformation is used here only as an operator to
obtain the orthogonal voltage phasors in a static reference
system, differently from other applications where it is used to
decouple the three-phase voltage system [11]. Then, the
application of the method does not depend on the transmission
line transposition scheme, i.e., it may be applicable to
untransposed lines as well.

x 10

x 10

0.8

V 0

V =
V

1.5

1.6

1.7

1.8
4

x 10

Fig. 3. Transient detection from a double phase to ground fault using Tdq0:
(a) Three-phase voltage signal; (b) Direct axis component Vd; (c) Difference
coefficients cdif; (d) Square of the difference coefficients [cdif]2.

IV. METHOD IMPLEMENTATION

The proposed method implementation is divided into five
steps: (a) voltage data acquisition using DFRs in both line
ends, (b) calculation of an orthogonal voltage phasor in a static

C. Orthogonal Voltage Phasors (Rotating Reference Frame)

This is the most important step in the proposed algorithm
because it makes possible the analysis of the aerial modes (V,
V) using a rotating reference frame, allowing the correct
transient detection. Parks transformation is applied to the
output voltage signals of Clarkes transformation to obtain the
voltage Vd, used to determine the fault location:
V d cos ( t + ) sin ( t + ) V .
V =

q sin ( t + ) cos ( t + ) V

(4)

Where:
is the angular power frequency;
t is the DFRs clock synchronized by GPS;
is the angle between Vd and the voltage phasor at phase A
(see Fig. 2b).

Voltage signals from both transmission line ends are used

to detect the first transient instants t11 and t21 at buses 1 and 2
(as shown in Fig. 1). The identification of t11 and t21 is shown
in Fig. 5.
Once t11 and t21 (in seconds), the line length and the
propagation velocity of aerial modes are known, (1) is used to
calculate the fault point location.
5

(a)
Three-phase
Voltage Signal - Bus 1 (V)

x 10
2

PhaseA

PhaseB

PhaseC

-1
-2
0.9

0.95

1.05

1.1
1.15
1.2
Sample Number

1.25

1.3

1.35

1.4
4

x 10

(b)
Three-phase
Voltage Signal - Bus 2 (V)

x 10
2

PhaseA

PhaseB

PhaseC

-1
-2
0.9

0.95

1.05

1.1
1.15
1.2
Sample Number

1.25

1.3

1.35

1.4
4

x 10

(c)
[cdif ] 2 coefficients - Bus 1

D. Detection of the Initial Transient Instants

In a two-terminals fault location method, [cdif]2 from both
line ends must be calculated. The analysis of the [cdif]2
coefficients allows the identification of the sample number in
which occur initial transients. So, to use (1) to estimate de fault
point, it is necessary to obtain the initial transient instants in
seconds. [cdif]2 with magnitude below a given preset threshold
are eliminated and, so, coefficients related to noise and low
frequency oscillation are ignored. Consequently, only [cdif]2
coefficients related to transients by the fault occurrence are
considered.
As shown in Fig. 4, first non-zero [cdif]2 occurs one sample
after the transient beginning. Finally, considering the sampling
frequency Fs used by DFRs, the initial transient instants ttransient
(in seconds) may be determined according to (5a) and (5b).

E. Fault Location Estimation

x 10

t11

Initial transient
instants detection

4
2
0
0.9

0.95
7

(d)
[cdif ] 2 coefficients - Bus 2

Apparently, the non alignment of Vd with the voltage phasor

of the phase A may pose a problem. As illustrated in Fig. 2, if
is greater than zero, a constant non-zero Vd signal is obtained
when the system is operating at normal conditions and, so,
transients may be wrongly detected. Here, the cdif coefficients
overcome that problem because, even getting constant nonzero Vd signals, each cdif coefficient would be nearly zero.
Therefore, the proposed method does not require alignment
between Vd component and the voltage phasor of the phase A
to detect transients correctly. Another good characteristic of
the method is that only current voltage samples are used in the
calculation of Vd coefficients, simplifying the implementation
and making possible the fault location almost immediately
after its occurrence. Other techniques as DWT and RDWT,
utilize more than one voltage or current sample of each phase
to calculate coefficients for the transient analysis of the
system, making fault location a more complex procedure.

x 10

1.05

1.1
1.15
1.2
Sample Number

1.25

1.1
1.15
1.2
Sample Number

1.25

1.3

1.35

1.4
4

x 10

t21

8
6
4
2
0
0.9

0.95

1.05

1.3

1.35

1.4
4

x 10

Fig. 5. Example of the initial transient instants detection in both terminals of

the transmission line through the [cdif]2 analysis: (a) Three-phase voltage
signals at Bus 1; (b) Three-phase voltage signals at Bus 2; (c) [cdif]2
coefficients related to Bus 1; (d) [cdif]2 coefficients related to Bus 2.

V. SIMULATION STUDIES
A. Simulation Model
Fig. 4. Detection of the initial transient instants through the [cdif]2 analysis.

If [c dif ( i ) ]2 = 0 t transient = {} ;

If c dif (i )

> 0 t transient = (i 1) F s ;

Where:
ttransient is the initial transient instant in seconds;
cdif is the difference coefficients calculated from Vd;
i is the sample number.

(5a)
(5b)

ATP [12] simulations of the 230 kV system presented in

Fig. 6 were performed to evaluate the proposed method. The
system parameters are shown in Table I. A constant
distributed-parameter transmission line model with a 50 s
time step was used. The line length is 500 km and the distance
d from bus 1 to the fault point is to be calculated. All
algorithm steps were implemented using the MODELS
language [13] and performed directly in the ATP one sample
at a time similar to what occurs in real time protection system,
with no need to analyze off line oscillographic records.

TABLE II
FAULT LOCATION SIMULATIONS FOR TRANSPOSED TRANSMISSION LINE
No of simulated
cases
960

Fault
type
AG

Fig. 6. Three-phase system considered in ATP simulations.

TABLE I
SYSTEM PARAMETERS
SOURCE A
SOURCE B
VG1 = 1,01410o pu VG 2 = 1,0000o pu

R0 = 1,445
G1

G2

= 1,445

G2

= 13,996 mH

G2

= 1,963

Z + = 0,054 + j0,527 km

= 14,982 mH

Y+

R0

G1

= 13,996 mH

L0

R+

G1

= 1,963

R+

L+

G1

= 14,982 mH

L0

L+

TRANSMISSION LINE
l = 500 km
Z0 = 0,236 + j1,035 km

G2

Y0

shunt

shunt

= 2,49 mho / km

960
960
960
960
960
960
960
960
960
9600

Reliable
simulations (%)
98.85%

949
949
960
960
960
960
960
960
960
9567

98.85%
98.85%
100.00%
100.00%
100.00%
100.00%
100.00%
100.00%
100.00%
99.66%

= 3,144 mho / km

B. Applied Disturbances
For validation purposes, situations where the transmission
line is transposed and untransposed were analyzed. For each
type of transposition scheme, 9600 different cases were run,
totaling 19200 different cases in which fault parameters were
changed: the fault resistance (5 to 95 with steps of 30 ),
the fault inception angle (0 to 180 with steps of 20) and the
distance from fault point to Bus 1 (20 km to 480 km with steps
of 20 km). Digital simulations were performed for phase to
ground, double phase, double phase to ground and three-phase
faults. The most adverse cases were considered, as high fault
resistance cases, fault inception angle near to zero (or at zero
crossing) and faults very close to the substations.
To perform such a high quantity of simulations, batch files
were created with Matlab to allow the automatic run of the
ATP files.
C. Simulation Results and Analysis
As shown in Table II and Table III, all results show the
reliability of the fault location method. It is important to point
out that the maximum expected error is a function of the
sampling rates, i.e., it is a hardware limitation. The travel time
of waves is not interpolated and, thus, the initial transient
instants are approximated to the nearly multiple of the time
step t used by DFRs. In this way, errors in the order of a half
time step can be introduced in the calculated initial transient
instants and are considered to be admissible errors. So, the
module of such admissible errors is calculated using (6).
e ( t c ) 2

BG
CG
AB
BC
CA
ABG
BCG
CAG
ABC
Total:

No of reliable
simulations
949

TABLE III
FAULT LOCATION SIMULATIONS FOR UNTRANSPOSED TRANSMISSION LINE

BG
CG
AB
BC
CA
ABG
BCG
CAG
ABC
Total:

For the used 20 kHz sampling frequency, the admissible

error is approximately 7.5 km in absolute value. This value is
used as the error threshold throughout the analysis of the
results. So, cases in which the fault location estimation
presents error above such threshold are classified as unreliable
cases.

960
960
960
960
960
960
960
960
960
9600

No of reliable
simulations
944

Reliable
simulations (%)
98.33%

947
950
954
960
954
955
960
959
960
9543

98.65%
98.96%
99.38%
100.00%
99.38%
99.48%
100.00%
99.90%
100.00%
99.41%

In Table IV is shown a general analysis of the simulations

in which transposed line and untransposed line cases are
considered.
In Table V, unreliable simulations cases are analyzed. As
expected, in most of unreliable simulations the errors are
related to high impedance fault cases and situations in which
fault inception angle is nearly to zero.
According to the results, less than 0.5% of 19200
performed cases diverged. Approximately 70% of unreliable
results are related to high impedance fault cases and only
about 6% are related to fault inception angle nearly to zero
crossing cases. So, the proposed method for fault location
presents low sensitivity in front of adverse cases related to
fault resistance and fault inception angle.

(6)

Where:
c is the speed of light ( 3.105 km/s).

No of simulated
cases
960

Fault
type
AG

TABLE IV
GENERAL RESULTS OF FAULT LOCATION SIMULATIONS
Considered
system

Transposed
Untransposed
Total:

No of
simulated
cases
9600

No of
unreliable
simulations
33

No of
reliable
simulations
9567

Reliable
simulations
(%)
99.66%

9600

57

9543

99.41%

19200

90

19110

99.53%

TABLE V
ANALYSIS OF UNRELIABLE SIMULATION RESULTS
Considered system
Case analysis
No of unreliable
simulations
No of cases related to
high impedance fault
(>60 )
No of cases related to fault
inception angle nearly to
zero crossing (< 10)
No of cases related to
unknown reasons

Transp.

Untransp.

Total:

33
(100.00%)
22
(66.67%)

57
(100.00%)
41
(71.93%)

90
(100.00%)
63
(70.00%)

0
(0.00%)

6
(10.53%)

6
(6.67%)

11
(33.33%)

10
(17.54%)

21
(23.33%)

The sampling frequency used by DFRs is a limitation of

fault location methods based on the theory of travelling waves.
In fact, increasing the sampling frequency Fs, the maximum
expected error decreases and a greater accuracy is obtained. In
order to address such considerations, some faults in the system
shown in Fig. 6 were performed using a transposed
transmission line model and different time steps to simulate
DFRs with sampling frequencies of 20 kHz, 100 kHz and 500
kHz. Simulation results for each sampling frequency are
shown in Table VI. The error can be compared to the
admissible error calculated from (6).
One can see that all estimated fault locations are within the
admissible errors and, consequently, the good accuracy, the
great simplicity and the reliability of the proposed technique
are attested.
TABLE VI
EFFECT OF THE SAMPLING FREQUENCY ON THE METHODS ACCURACY
FT

FP
(km)

Fs=20 kHz
FL
(km)

Fs=100 kHz

(km)

FL
(km)

(km)

The proposed method was applied to transmission lines

taking into account the influence of fault resistance, voltage
inception angle and the transmission line transposition scheme.
An analysis of the effect of the sampling frequency on the
methods accuracy was performed. The algorithm was
implemented in ATP using the MODELS language. Batch files
were created with Matlab to allow the automatic run of the
ATP files. A total of 19200 EMTP simulations for a 230 kV
transmission line were performed. In more than 99% of the
simulated cases, the proposed algorithm presented errors
below the admissible value.
Finally, the proposed method is very simple, fairly accurate,
do not require phasor estimation and may be used as an
additional routine in transmission lines protective relays.
VII. ACKNOWLEDGMENT
The authors would like to thank the reviewers for
invaluable suggestions.
VIII. REFERENCES
[1]

[2]

[3]

[4]

Fs=500 kHz
FL
(km)

(km)

AG

28

23.30

-4.70

27.69

-0.31

28.27

0.27

BG

78

81.80

3.80

77.41

-0.59

78.29

0.29

CG

129

125.68

-3.32

130.07

1.07

129.19

0.19

AB

166

169.56

3.56

165.17

-0.83

166.05

0.05

BC

205

198.81

-6.19

204.66

-0.34

204.95

-0.05

CA

244

242.69

-1.31

244.15

0.15

243.86

-0.14

ABG

311

315.82

4.82

311.43

0.43

310.84

-0.16

BCG

393

388.95

-4.05

393.33

0.33

392.75

-0.25

CAG

432

432.82

0.82

431.36

-0.64

431.65

-0.35

ABC

472

476.70

4.70

472.31

0.31

471.73

-0.27

[5]

[6]

[7]

[8]

Where: FT = Fault Type; FP = Fault Point; FL = Fault Location; = Error;

Fs = Sampling Frequency.
[9]

VI. CONCLUSIONS
In this paper an approach for fault location on transmission
lines monitored in two terminals was presented. Only the
current voltage samples are used to detect transients and all
three phases are monitored by the direct axis voltage signal Vd
obtained through Parks transformation (Tdq0). The method
automatically determines the fault location point immediately
after the fault occurrence.

[10]

[11]

[12]
[13]

H. W. Dommel, J. Michels, High-speed Relaying Using Travelling

Wave Transient Analysis. IEEE Conference, Paper No. A78, pp. 214219, January/February 1978.
M. Gilany, DK. Ibrahim, E. S. T. Eldin, "Traveling-Wave-Based FaultLocation Scheme for Multiend-Aged Underground Cable System,"
IEEE Transactions on Power Delivery, vol. 22, pp. 82-89, January
2007.
P. F. Gale, P. A. Crossley, Xu Bingyin, Ge Yaozhong, B. J. Cory, J. R.
J. Barker, "Fault Location Based on Travelling Waves," Fifth
International Conference on Developments in Power System
Protection, pp. 54-59, 1993.
F. B. Costa, B. A. Souza, N. S. D. Brito, K. M. Silva, "Discrete Wavelet
Transform applied to the disturbance diagnosis," (in Portuguese),
Brazilian Symposium of Electrical Systems, Belm, Brazil, 2010.
D. Lin, P. Jun, S. Wenxia, T. Jun, Z. Jun, "Fault Location for
Transmission Line Based on Traveling Waves using Correlation
Analysis Method," International Conference on High Voltage
Engineering and Application, Chongqing, China, November 9-13,
2008.
Z. Feng, L. Jun, Z. Li, Y. Zhihao, "A New Fault Location Method
Avoiding Wave Speed and Based on Traveling Waves for EHV
Transmission Line," Electric Utility Deregulation and Restructuring
and Power Technologies-DRPT2008, Nanjing, China, April 6-9, 2008.
X. Zheng, X. Li, J. Ding, Z. Duan, "Study on Impedance-traveling Wave
Assembled Algorithm in One-terminal Fault Location System for
Transmission Lines," Electric Utility Deregulation and Restructuring
and Power Technologies-DRPT2008, Nanjing, China, April 6-9, 2008.
M. Aurangzeb, P. A. Crossley, P. F. Gale, "Fault Location on a
Transmission Line Using High Frequency Travelling Waves Measured
at a Single Line End," IEEE Power Engineering Society Winter
Meeting, vol. 4, pp. 2437 - 2442, August 06, 2002.
R. H. Park, "Two-Reaction Theory of Synchronous Machines,
Generalized Method of Analysis Part I," Winter Convention of the A.
I. E. E., pp. 716-730, New York, January 28-February 1, 1929.
G. Benmouyal, J. Mahseredjian, "A Combined Directional and Faulted
Phase Selector Element Based on Incremental Quantities," IEEE
Transactions on Power Delivery, vol. 16, issue: 4, pp. 478-484, October
2001.
L. M. Wedepohl, "Application of Matrix Methods to the Solution of
Travelling-Wave Phenomena in Polyphase Systems," Proceedings of the
IEE, vol. 110, no. 12, pp. 2200-2212, December 1963.
Leuven EMTP CENTER, ATP - Alternative Transients Program: Rule
Book. Heverlee, Belgium, July 1987.
L. Dube, "Models in ATP: Language Manual," [S.l.], 1996.