This article has been accepted for inclusion in a future issue of this journal.
Content is final as presented, with the exception of pagination.
IEEE TRANSACTIONS ON POWER DELIVERY
Unsynchronized Fault Location Based on the
Negative-Sequence Voltage Magnitude for
Double-Circuit Transmission Lines
Behnam Mahamedi and Jian Guo Zhu, Senior Member, IEEE
AbstractThis paper describes a new approach to fault location
for double-circuit transmission lines based on only the voltage data
of both ends of the faulted circuit. The ratio between the magni-
tudes of negative-sequence voltages measured at both ends of the
faulted circuit is utilized to estimate the fault location. Since only
the magnitudes are used, the data of both ends are not required to
be synchronized, which removes any concern about data synchro-
nization. Moreover, since only the voltage data are required, the er-
rors caused by current transformers can be avoided. The proposed
method can effectively locate the single-phase-to-ground, double-
phase-to-ground, and phase-to-phase faults disregarding the fault
resistance and prefault conditions and without any need for fault
classification and phase selection. Unlike the iterative methods, the
proposed method is fully analytical and does not cause much com-
puting burden to the line relays. The accuracy and practicality of
the proposed method make it an attractive function to implement
in numerical relays.
Index TermsDouble-circuit transmission line, fault location,
negative-sequence reactance, negative-sequence voltage magni-
tude, unsynchronized measurement.
NOMENCLATURE
and Negative-sequence reactance behind
relays and , respectively.
Negative-sequence reactance of a
transmission line.
and Negative-sequence voltage measured by
relays and , respectively.
Negative-sequence voltage at the fault
point.
and Negative-sequence capacitance and
capacitive reactance of a transmission
line, respectively.
and Equivalent capacitance at the sending and
receiving ends, respectively.
Manuscript received June 09, 2013; revised October 30, 2013; accepted De-
cember 04, 2013. Paper no. TPWRD-00660-2013.
B. Mahamedi was with the School of Electrical, Mechanical and Mechatronic
Systems, University of Technology Sydney, Ultimo NSW 2007, Australia. He
is now with the Iran Grid Management Company, Tehran IR-15175-648, Iran
(e-mail: behnam.mahaamedi@gmail.com).
J. G. Zhu is with the School of Electrical, Mechanical and Mechatronic Sys-
tems, University of Technology Sydney, Ultimo NSW 2007, Australia (e-mail:
jianguo.zhu@uts.edu.au).
Digital Object Identifier 10.1109/TPWRD.2013.2294972
0885-8977 2014 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
1
Ratio between the magnitudes of
negative-sequence voltages at the local
and remote ends.
Fault location in the faulted circuit of the
double-circuit transmission line in per
unit.
I. INTRODUCTION
D OUBLE-CIRCUIT transmission lines have been adopted
in modern power systems to increase the transmission
capacity and to enhance reliability and security. However,
protecting double-circuit lines, in particular locating faults
on them, are always challenging due to the mutual coupling
between zero-sequence components of the two circuits.
Whether they are for single-circuit or double-circuit transmis-
sion lines, there are two types of fault-location methods known
as one-end-based and two-end-based ones. The one-end-based
methods [1], [2] may suffer from errors due to the variations
of source impedance, fault incidence angle, and loading condi-
tions while their major merit is that they need data of only one
end [3]. To overcome the shortfalls of one-end-based methods,
two-end-based methods were introduced. They can be classified
into unsynchronized and synchronized ones. Unsynchronized
methods do not need the data of both line ends to be synchro-
nized while data synchronization is the first step in the synchro-
nized methods.
Up till now, several methods have been proposed for fault
location on double-circuit transmission lines, most of which re-
quire data of both voltage and current [4][11]. The magnetic
saturation of current transformers (CTs) always presents chal-
lenges of having reliable current data and, thus, a strong desire
for possible approaches based on only voltage data exists.
Chen et al. [4] presented a fault-location method based on the
distributed line model. This method requires the phasor mea-
surement units (PMUs) for synchronized data and continuous
monitoring of the line under normal operation. Funabashi et al.
[5] presented two algorithms for multiterminal double-circuit
transmission lines based on impedance calculation and cur-
rent diversion ratio. The synchronization for voltages and cur-
rents at all terminals is required. Kang et al. [6] developed a
fault-location algorithm for untransposed double-circuit trans-
mission lines based on the lumped parameter model of trans-
mission lines by ignoring the line shunt capacitances. This may
cause significant errors in long transmission lines. Besides, fault
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

2 IEEE TRANSACTIONS ON POWER DELIVERY

classification is required for fault location. Izykowski et al. [7]

proposed a method which requires the voltage and current data
of faulted and healthy circuits, but the measurement from the
healthy circuit is not always available.
Apostolopoulos and Korres [8] put forward an algorithm
based on the positive-sequence current without utilizing the
line parameters, which requires prefault and postfault data.
Kang and Liao [9] proposed a method using synchronized Fig. 1. Double-circuit transmission line protected by the distance relays of
sending and receiving ends, and .
current phasors. The data synchronization of various busbars,
however, requires PMUs to be installed on these busbars.
Apostolopoulos and Korres [10] proposed an iteration-based
method, but multiple iterations can cause heavy computing
burden on the relays. Song et al. [11] described a time-domain
fault-location algorithm for parallel transmission lines using
currents of both ends based on a differential component net.
The sampling frequency required for the method is much higher
than the practical one that is around 1 kHz [12].
Assuming both relays are provided with the source
impedance behind them, which are available at load dis-
patch centers [13], this paper presents a new fault-location
method for double-circuit transmission lines using the mag-
nitudes of negative-sequence voltages measured at both ends
of the faulted circuit. The faulted circuit is assumed to be
known based on distance relays action and circuit-breaker (CB)
status. In this method, fault classification is not required to
estimate fault location. This is a specific feature in comparison
to the fault locators that need to first classify the fault type.
One interesting attribute of the proposed method is that the
fault-location accuracy still remains high for long transmission
Fig. 2. Negative-sequence circuit of the double-circuit transmission line for an
lines even when the lumped model is employed. It is true unbalanced fault occurring at location (in per unit) from relay .
since the proposed method uses only the voltage data and the
ratio between voltages of both ends. Moreover, the proposed
method is fully analytical even when the shunt capacitances negative-sequence circuit can thus be modeled as shown in
are considered. The accuracy of the method, together with its Fig. 2.
practicality, makes it an attractive option to implement in the Converting the delta connection formed by the healthy circuit
relays as a fault-location function. The proposed method is a reactance and the source reactances to Y connection, the nega-
modified and extended version of the one proposed in [14]. tive-sequence circuit can be represented in a new form as shown
The rest of this paper is organized as follows. Section II in Fig. 3, where , and can be obtained from
presents a detailed description of the proposed method, and Fig. 2 as
Section III investigates the effects of the line shunt capacitances
on the derived equations. Section IV is devoted to studying the (1)
sensitivity of the proposed method to the parameters which are
required for fault location. Section V presents the numerical (2)
simulation to demonstrate the proposed method. The simulation
results were obtained by using the steady-state data of postfault (3)
analysis carried out in MATLAB/SIMULINK. Finally, conclu-
and
sions are drawn in Section VI.
(4)
II. PROPOSED METHOD
Consider a double-circuit transmission line as shown in From Fig. 3, the following equations can be derived:
Fig. 1, which is protected by distance relays and . The
relays are fed by only voltage transformers to emphasize that
the proposed method is based on only the voltage data.
In transmission systems, the resistances are negligible in
comparison to the reactances [15], and the mutual coupling
between the negative-sequence components of each circuit is
also negligible [16]. If an unbalanced fault occurs at distance
(5)
[per unit (p.u.)] from relay in one circuit of the line, the
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
MAHAMEDI AND ZHU: UNSYNCHRONIZED FAULT LOCATION BASED ON THE NEGATIVE-SEQUENCE VOLTAGE MAGNITUDE
Fig. 3. Negative-sequence circuit of the double-circuit transmission line after
converting the delta connection to the Y connection.
Fig. 4. Line shunt capacitances considered in the negative-sequence circuit of
the double-circuit transmission line.
(6)
3
where is the voltage between the fault point and the star
point of the Y connection in the negative-sequence circuit.
Considering only the phasor magnitudes and dividing (5) by
(6) yields
(7)
Solving (7) in terms of leads to
(8)
As a result, knowing (the ratio between the magnitudes of
negative-sequence voltages at the sending, S, and receiving, R,
ends of the faulted circuit) and the negative-sequence reactances
of the sources and that of the line, the fault location in a double-
circuit transmission line can be estimated by using (8).
Here, it can be argued that the same results can also be ob-
tained by using the zero-sequence circuit. The following out-
lines the reasons why the negative-sequence circuit rather than
the zero-sequence one was chosen. The first and an evident one
is that the zero-sequence circuit does not exist in the case of un-
grounded faults. The second is that there is mutual coupling be-
tween zero-sequence components of two circuits in double-cir-
cuit transmission lines. The third is that the zero-sequence pa-
rameters are not as accurate as the negative-sequence ones due
to some unknown parameters that contribute to the zero-se-
quence circuits. Finally, it should be mentioned that the zero-se-
quence resistances cannot be assumed negligible in comparison
to the zero-sequence reactances for developing a fault-location
method since it might result in a large error. This assumption
may, however, work well in other applications, such as fault de-
tection and classification [17].
III. CONSIDERATION OF SHUNT CAPACITANCES
To consider the effects of the line shunt capacitances on the
proposed method, a PI model for the double-circuit transmission
line is used. The PI model is quite accurate to obtain steady-
state results where the shunt capacitances should be considered
[18]. Based on the PI model, if an unbalanced fault occurs at
location m (in p.u.) from relay in one circuit of the double-
circuit transmission line, the negative-sequence circuit will be
as shown in Fig. 4.
The capacitances of the healthy circuit are parallel to those of
the faulted circuit at the sending and receiving ends. The equal
capacitances at the sending and receiving ends will then be ob-
tained as follows:
(9)
(10)
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

4 IEEE TRANSACTIONS ON POWER DELIVERY

The equal capacitances given in (9) and (10) are parallel with TABLE I
the source reactances and , respectively. Thus, two new DIFFERENT FAULT CASES CONSIDERED FOR EVALUATION STUDIES
source reactances and will be formed by the parallel
combination of the source reactances and the equal capacitances
as

(11)

(12)

Substituting and as and in (8), one

obtains as
Similarly, sensitivity factors can be obtained by using (13)
when line shunt capacitances are considered.
(13)
V. EVALUATION AND SIMULATION RESULTS
A 400-kV, 50-Hz double-circuit transmission line of 200-km
IV. SENSITIVITY ANALYSIS length illustrated in Fig. 1 was chosen to evaluate the accuracy
of the proposed method. The parameters of the transmission line
It is a worthy practice to study the sensitivity of the proposed and those of the sources are given in the Appendix. Since the
method to the required parameters. As shown in (8) and (13), the proposed method is based on the voltage data, capacitive voltage
source parameters and and line parameters and transformers (CVTs) were modeled in detail as per data given
are required for estimation of fault location. The line reac- in [19]. Also, the fault inception time was chosen as 0.1
tance and capacitance can be accurately measured but the source s. The negative-sequence voltage magnitude was calculated by
reactances may not be obtained as accurately as the line parame- the discrete Fourier transform (DFT) as one of the most common
ters. Therefore, the sensitivity of fault-location estimation to the approaches.
source reactances should be studied. Here, one-at-a time (OAT) Different fault cases presented in Table I were considered to
normalized local sensitivity analysis is adopted, which is a basic evaluate the proposed method.
approach in sensitivity analysis.
It is assumed that the expected value of (the ratio of nega- A. Accuracy Evaluation
tive-sequence voltage magnitude at the sending-end relay to the Figs. 513 show the estimation errors of fault location for the
one at the receiving-end relay) is first measured while the fault different fault cases of Table I. The errors were calculated by
location varies along the line. Having known the variation of ,
the sensitivity of fault location to the aforementioned parame- Estimation error%
ters will be obtained as follows. Actual location - Estimated location
Taking the derivative of (8) in terms of yields the sen- Total line length
sitivity factor to the source reactance behind the sending-end (16)
relay as
To provide a better vision of the accuracy of the method, here
it is preferred to present the real value of the percentage errors,
not their absolute values. The absolute values of the estimation
errors Error (%) have been used to provide the maximal and
mean values of the errors shown in Table II.
The mean values for each fault case have been calculated by
(14) averaging out all of the absolute errors obtained at different fault
locations (every 1 km along the line).
The sensitivity factor to the source reactance behind the re- The results confirm that the proposed fault-location method
ceiving-end relay is obtained by taking the derivative of (8) in is independent of the fault type and fault resistance. This can
terms of as also be inferred from the discussion presented in Sections II
and III. The small difference in the estimation errors when the
fault type and fault resistance vary can be attributed to the errors
generated by the instrument transformers (CVTs) and numerical
calculation. It can be seen that the accuracy of the method will
be quite high if the required parameters are close enough to their
(15) real values.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

MAHAMEDI AND ZHU: UNSYNCHRONIZED FAULT LOCATION BASED ON THE NEGATIVE-SEQUENCE VOLTAGE MAGNITUDE 5

Fig. 5. Estimation errors for Case 1 based on the lumped and PI model. Fig. 8. Estimation errors for Case 4 based on the lumped and PI model.

Fig. 9. Estimation errors for Case 5 based on the lumped and PI model.
Fig. 6. Estimation errors for Case 2 based on the lumped and PI model.

Fig. 10. Estimation errors for Case 6 based on the lumped and PI model.
Fig. 7. Estimation errors for Case 3 based on the lumped and PI model.

real values and . To do this, it was assumed that the

B. Sensitivity Evaluation source reactances used for estimation are given either 110%
Fig. 14 shows the measured value of for different fault loca- or 90% of their real values. In this sense, four possible cases
tions in the simulated system. The normalized sensitivity factors for the source reactances emerged as 1) 1.1 X , 1.1 X ;
to and are, respectively, represented by and and 2) 0.9 X , 0.9 X ; 3) 1.1 X , 0.9 X ; and 4)
shown in Figs. 15 and 16. It can be seen that for fault locations 0.9 X , 1.1 X . Figs. 1720 show the estimation errors
of more than 0.2 p.u., both sensitivity factors are less than 1, for each case. As shown, even though both source reactances
which indicates a low sensitivity of the method to the variation deviate 10% from their real values at the same time, the esti-
of source reactances. For instance, and are, respectively, mation errors still fall in a reasonable range with a maximum
calculated as 0.66 and 0.75 at 0.3. If is given with error of 5%.
10% tolerance, the fault location will be estimated in the range
of [0.280, 0.320]. For with a tolerance of 10%, the fault VI. CONCLUSION
location will be estimated in the range of [0.277, 0.323]. In case A new fault-location method for double-circuit transmission
of a fault location farther from relay , the accuracy will even lines is proposed in this paper. It is a steady-state-based method
be higher. based on the ratio of negative-sequence voltage magnitudes at
The aforementioned sensitivity factors were obtained based both ends of the line. The evaluation results are encouraging
on OAT sensitivity analysis illustrated in Section IV. Here, it and have shown the practicality of the proposed method for im-
is worth investigating how much the estimation errors would plementation in real applications. The proposed concept can be
be if both source reactances simultaneously deviate from their readily developed for single-circuit transmission lines as well.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

6 IEEE TRANSACTIONS ON POWER DELIVERY

Fig. 11. Estimation errors for Case 7 based on the lumped and PI model.
Fig. 14. Variation of for different fault locations.

Fig. 12. Estimation errors for Case 8 based on the lumped and PI model.

Fig. 15. Sensitivity factor to the source reactance behind the sending-end relay
for different fault locations based on the lumped model.

Fig. 13. Estimation errors for Case 9 based on the lumped and PI model.

TABLE II
MAXIMAL AND MEAN VALUES OF ABSOLUTE ERRORS FOR EACH FAULT CASE
BASED ON THE LUMPED AND PI MODELS

Fig. 16. Sensitivity factor to the source reactance behind the receiving-end
relay for different fault locations based on the lumped model.

The proposed method considers the lumped parameter and

PI models of a transmission line. It was found that even
by considering the lumped parameter model the estimation
errors are small for a long transmission line. The reason
is that the proposed method does not use the current data
and requires only the ratio between voltage magnitudes.
Other existing methods may fail for long transmission lines
if they do not consider shunt capacitances.
The major advantages of the proposed method can be sum- Most of the existing methods require both voltage and cur-
marized as follows. rent data at either one or both ends of a transmission line.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

MAHAMEDI AND ZHU: UNSYNCHRONIZED FAULT LOCATION BASED ON THE NEGATIVE-SEQUENCE VOLTAGE MAGNITUDE 7

Since the proposed method needs only the voltage data at

Fig. 17. Estimation errors for a deviation of 10% in source reactances (case:
1.1 X and 1.1 X ).
both ends of the line, it eliminates the errors produced by
the CT measurement.
One of the attributes of the proposed method is that it is
independent of the fault resistance and fault type. The small
differences in the estimation errors for different fault types
and fault resistances are due to CVT measurement affected
by the transients which contribute to computing errors.
Another interesting feature of the proposed method
is that it requires only the magnitudes of negative-se-
quence voltages, which completely removes the need for
synchronization.
The proposed method is a purely analytical method, which
effectively eliminates the computing burden to the relay.
The only type of fault that cannot be located by the pro-
posed method is the three-phase fault which hardly occurs
on transmission lines.

APPENDIX
The parameters of the simulated system are given as follows.
The positive-, negative- and zero-sequence impedances be-
hind relay are

Fig. 18. Estimation errors for deviation of 10% in source reactances (case:
0.9 X and 0.9 X ).
The voltage of the source behind relay is

The positive-, negative- and zero-sequence impedances be-

hind relay are

The voltage of source behind relay is

Fig. 19. Estimation errors for deviation of 10% in source reactances (case:
1.1 X and 0.9 X ).
The distributed model parameters of the double-circuit trans-
mission line:
The positive-, negative-, and zero-sequence impedances are

The mutual coupling impedance between the zero-sequence

circuits is

The positive-, negative-, and zero-sequence capacitances are

Fig. 20. Estimation errors for deviation of 10% in source reactances (case:
0.9 X and 1.1 X ).
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
8 IEEE TRANSACTIONS ON POWER DELIVERY
The mutual coupling capacitance between the zero-sequence
circuits is
REFERENCES
[1] T. Kawady and J. Stenzel, A practical fault location approach for
double circuit transmission lines using single end data, IEEE Trans.
Power Del., vol. 18, no. 4, pp. 11661173, Oct. 2003.
[2] G. Song, J. Suonan, and Y. Ge, An accurate fault location algorithm
for parallel transmission lines using one-terminal data, Elect. Power
Energy Syst., vol. 31, no. 23, pp. 124129, 2009.
[3] J.-A. Jiang, J.-Z. Yang, Y.-H. Lin, C.-W. Liu, and J.-C. Ma, An adap-
tive PMU based fault detection/location technique for transmission
lines part I: Theory and algorithms, IEEE Trans. Power Del., vol. 15,
no. 2, pp. 486493, Apr. 2000.
[4] C. S. Chen, C. W. Liu, and J.-A. Jiang, A new adaptive PMU based
protection scheme for transposed/untransposed parallel transmission
lines, IEEE Trans. Power Del., vol. 17, no. 2, pp. 395404, Apr. 2002.
[5] T. Funabashi, H. Otoguro, Y. Mizuma, L. Dube, and A. Ametani, Dig-
ital fault location for parallel double-circuit multi-terminal transmis-
sion lines, IEEE. Trans. Power. Del., vol. 15, no. 2, pp. 531537, Apr.
2000.
[6] S. Kang, Y. Ahn, Y. Kang, and S. Nam, A fault location algorithm
based on circuit analysis for untransposed parallel transmission lines,
IEEE Trans. Power Del., vol. 24, no. 4, pp. 18501856, Oct. 2009.
[7] J. Izykowski, E. Rosolowski, and M. M. Saha, Locating faults in par-
allel transmission lines under availability of complete measurements at
one end, Proc. Inst. Elect. Eng., Gen., Transm. Distrib., vol. 151, no.
2, pp. 268273, Mar. 2004.
[8] C. A. Apostolopoulos and G. N. Korres, A novel fault-location al-
gorithm for double-circuit transmission lines without utilizing line pa-
rameters, IEEE. Trans. Power. Del., vol. 26, no. 3, pp. 14671478,
Jul. 2011.
[9] N. Kang and Y. Liao, Double-circuit transmission-line fault location
utilizing synchronized current phasors, IEEE. Trans. Power. Del., vol.
28, no. 2, pp. 10401047, Apr. 2013.
[10] C. A. Apostolopoulos and G. N. Korres, Accurate fault location al-
gorithm for double-circuit series compensated lines using a limited
number of two-end synchronized measurements, Elect. Power Energy
Syst., vol. 42, no. 1, pp. 495507, 2012.
[11] G. Song, J. Suonan, Q. Xu, P. Chen, and Y. Ge, Parallel transmission
lines fault location algorithm based on differential component net,
IEEE Trans. Power Del., vol. 20, no. 4, pp. 23962406, Oct. 2005.
[12] I. Zamora, A. J. Mazn, V. Valverde, E. Torres, and A. Dysko, Power
quality and digital protection relays, presented at the Int. Conf. Renew.
Energies Power Qual., Barcelona, Spain, 2004.
[13] S. M. Brahma and A. A. Girgis, Fault location on a transmission line
using synchronized voltage measurements, IEEE Trans. Power Del.,
vol. 19, no. 4, pp. 16191622, Oct. 2004.
[14] Y. Liao, Fault location utilizing unsynchronized voltage measure-
ments during fault, Elect. Power Compon. Syst., vol. 34, no. 12, pp.
12831293, 2006.
[15] H. Saadat, Power System Analysis, 2nd ed. New York: McGraw-Hill,
2002.
[16] A. Apostolov, D. Tholomier, S. Sambasivan, and S. Richards, Protec-
tion of double circuit transmission lines, presented at the 60th Annu.
Conf. Protect. Relay Eng., College Station, TX, USA, Mar. 2007.
[17] B. Mahamedi and J. G. Zhu, Fault classification and faulted phase
selection based on the symmetrical components of reactive power for
single-circuit transmission lines, IEEE Trans. Power Del., vol. 28, no.
4, pp. 23262332, Oct. 2013.
[18] O. Hevia, Alternative Transients ProgramComparison of
transmission line models. [Online]. Available: www.iitree-
unlp.org.ar/caue/Archivos/emodlin.pdf
[19] D. Tziouvaras et al., Mathematical models for current, voltage and
coupling capacitor voltage transformers, IEEE Trans. Power Del., vol.
15, no. 1, pp. 6272, Jan. 2000.
Behnam Mahamedi was born in Tehran, Iran, in
1982. He received the B.Sc. degree in electrical
engineering from Iran University of Science and
Technology (IUST), Tehran, in 2005 and the M.Sc.
degree in electrical engineering from Shahid Be-
heshti University (SBU), Tehran, in 2010.
From 2012 to 2013, he was a Researcher with the
University of Technology Sydney (UTS), Sydney,
Australia, focusing on power system fault detection
systems. Since 2008, he has been a Protection
Engineer with the Iran Grid Management Company
(IGMC), Tehran. His research interests are power system protection, power
system transients, and wavelet-transform applications in power systems.
Jian Guo Zhu (SM03) received the B.E. degree in
electrical engineering from Jiangsu Institute of Tech-
nology, Zhenjiang, China, in 1982, the M.E. degree
in electrical engineering from Shanghai University of
Technology, Shanghai, China, in 1987, and the Ph.D.
degree in electrical engineering from the University
of Technology Sydney (UTS), Sydney, Australia, in
1995.
Currently, he is a Professor of Electrical En-
gineering and Head of the School of Electrical,
Mechanical, and Mechatronic Systems, UTS. His
research interests include electromagnetic and magnetic properties of mate-
rials, electrical machines and drives, power electronics, and renewable energy
systems.