Electric Power Systems Research 220 (2023) 109302

Contents lists available at ScienceDirect

Electric Power Systems Research

journal homepage: www.elsevier.com/locate/epsr

A single-ended traveling wave based fault location method using DWT in

hybrid parallel HVAC/HVDC overhead transmission lines on the
same tower
Mohammad Fayazi , Mahmood Joorabian *, Alireza Saffarian , Mehdi Monadi
Department of Electrical Engineering, Faculty of Engineering, Shahid Chamran University of Ahvaz, Ahvaz 61357-85311, Iran

A R T I C L E I N F O A B S T R A C T

Keywords: In this paper, a single ended traveling wave based fault location method is presented for Hybrid Parallel HVAC/
Hybrid parallel HVAC/HVDC overhead HVDC Overhead Transmission Lines (HPOTLs) on the same tower using Discrete Wavelet Transform (DWT). In
transmission line HPOTLs, the AC and DC transmission lines are adjacent to each other from the sending to the receiving end of the
AC and DC faults
lines and there is electromagnetic coupling between them. Fault detection and classification algorithms in
Intersystem faults
Fault location
HPOTLs are presented in another paper. Therefore, in order to summarizing, in this paper, it is assumed that the
Discrete Wavelet Transform (DWT) fault is detected in the HVAC or/and HVDC lines and faulty line is identified. HPOTL faults include AC faults, DC
Traveling wave (TW) faults and AC/DC intersystem faults. AC and DC voltage signals on sending end of the transmission line with a
Signal processing sampling frequency of 100 kHz are used for fault location in this paper. In order to minimize the noise effect, the
squared wavelet transform coefficients (WTC2) is used for proposed fault location method. Finally, Bewley-
Lattice diagrams are checked for the traveling wave pattern and locate the fault. At the end of paper, it has
been shown that the proposed method has high accuracy, noise immunity, robustness against fault inception
angles, and high impedance faults.

1. Introduction next step which is of considerable importance is the precise and fast fault
location to reduce the downtime and repair costs and hence improve the
Nowadays, HVDC transmission lines are widely used in transmission availability and reliability of the transmission system. Therefore, some
networks due to many advantages such as higher power transmission papers on fault location in AC and DC transmission lines have been
capacity over long distances, higher reliability and lower transmission researched and methods have been provided to determine the location
losses than HVAC transmission lines. For the reasons mentioned, in some of the fault.
existing transmission lines in the world, to increase the transmission Fault location methods in transmission lines are usually classified
capacity of HVAC transmission lines, multi-circuit AC lines are con­ into three groups: impedance-based methods, traveling wave-based
verted into hybrid parallel HVAC/HVDC transmission lines [1]. In some methods and learning-based methods [5,6]. The impedance-based
other cases, due to environmental reasons and reduction of transmission methods are simple and low costs. In this methods, the synchronized
line Right-of-Way (ROW), AC and DC transmission lines are transferred voltage and current phasors obtained from phasor measurement units or
in parallel and adjacent to each other. In all these cases, due to the intelligent electronic devices are employed to identify the fault line
proximity of AC and DC transmission lines and the induction of voltage section and locate fault positions [7]. Other fault location methods,
and current components in both transmission lines, the performance of traveling wave-based methods and learning-based methods are widely
each line in different conditions affects the other line, especially its used in both AC and DC transmission lines. Therefore, nowadays, the use
protection systems and fault location algorithms [2–4]. This effect can of these methods has received much attention in the papers. The trav­
be caused by electromagnetic coupling or direct galvanic contact. Fault eling wave-based methods have high accuracy and the results are not
location in HPOTLs is an issue that will become more important with the affected by the parameters, such as ground resistance or loading con­
growth of HPOTLs in the transmission network. ditions [8]. The detection of traveling waves is done using methods such
Following the successful detection and classification of the fault, the as correlation analysis, maximum likelihood estimation (MLE), adaptive

* Corresponding author.
E-mail address: [email protected] (M. Joorabian).

https://doi.org/10.1016/j.epsr.2023.109302
Received 27 December 2022; Received in revised form 28 February 2023; Accepted 6 March 2023
Available online 22 March 2023
0378-7796/© 2023 Published by Elsevier B.V.
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

filtering, mathematical morphology, empirical mode decomposition But fault location in these lines, especially the AC/DC intersystem faults,
(EMD) and wavelet transform (WT). is not presented in the papers. The purpose of this paper is to locate the
Another classification of fault location methods in transmission lines fault in HPOTLs (AC and DC transmission lines on the same tower). In
is based on measurements they employ at receiving and/or sending ends this paper, in addition to fault location in AC and DC transmission lines
of the transmission lines [9]. Therefore, these methods can be classified separately, fault location for AC/DC intersystem faults is also done with
into two categories, single-ended and multi-ended methods. High reli­ a similar method.
ability, no need for data communication and data synchronization, less In all the simulations performed, electromagnetic coupling between
complexity and lower costs are the attractions of fault location methods AC and DC transmission lines is considered. Simulations are carried out
based on single-ended data [10]. Since the method used in this paper is by using PSCAD, and MATLAB (MATLAB Wavelet Toolbox) is used for
based on traveling waves, in the following, papers with the method of data analysis and signal processing. The remainder of this paper is
fault location based on traveling waves have been reviewed. organized as follows: Section 2 explains the fault location principles and
The difference between the methods in the papers that have used gives a brief introduction of DWT, traveling wave theory and Trans­
traveling wave-based methods for fault location is the type of signal used mission line faulty half identification method. Section 3 describes the
(voltage or current) and sampling frequency. Most of the fault location HPOTL modeling, and Section 4 discusses the simulation results.
schemes are based on voltage traveling wave propagation on the Sensitivity analysis of the proposed fault location method are presented
transmission line [11]. The use of discrete wavelet transform (DWT) in in section 5. In section 6, the discussion and comparison of the results
fault location in traveling wave based methods is first presented in [12]. between different papers and the proposed method in the paper are
Double-ended traveling wave-based fault location method using DWT in presented and finally Section 7 concludes the paper.
underground cables is presented in [13]. A traveling-wave based fault
location method for two terminal HVAC transmission lines using un­ 2. Fault location principles
synchronized current measurements from intelligent electronic devices
(IEDs) recorded at both ends is presented in [14]. In this paper, first and This section presents the principles of fault location in HPOTLs based
second traveling wave arrival time and propagation velocity of the line on sending end voltage signals of transmission line. In the proposed
are used for fault location. High sampling frequency is one of the method, the voltage signals of AC and DC transmission lines are sampled
weaknesses of the proposed method. In the single-ended method, the on sending end of transmission lines at the sampling rate of 100 kHz.
fault transients, which is reflected from the fault point and arrive at the After AC and DC voltage sampling, the modal components are extracted
relay terminals, produce a highly correlated signal for a delay time equal from the three-phase voltages of the AC line and the positive and
to twice the traveling time of the transients to the fault location. This negative pole voltages of the DC line. In this paper, the aerial mode of AC
time can be then used to find the distance from the relay to the fault and line mode (1-mode) of DC voltage is used to calculate the location of
location [11]. In [15,16], the single-ended traveling wave based method AC, DC and AC/DC intersystem faults in transmission line. It should also
has been used to fault location in AC and DC transmission lines. In [16], be mentioned that in order to improve the estimation of the fault loca­
a sampling frequency of 1 MHz was used and high impedance faults tion in AC/DC intersystem faults that occur between AC and DC lines,
were not checked. [17] Has used the dominant natural frequency in the the aerial mode of the voltage of both transmission lines (AC and DC) are
spectrum analysis of current traveling waves to locate faults in DC used simultaneously and the fault location is calculated based on the
transmission lines. In [11], a comprehensive review on fault location in averaging between them. The modal voltage signals are decomposed
AC transmission lines based on current traveling waves using single-end using the db4 wavelet for AC voltage (scale 1) and Haar wavelet (db1)
and double-end data is presented. The sampling frequency used in this for DC voltage signals (scale 2). The squared wavelet transform co­
paper is 1.25 MHz, which is considered as a weakness for the proposed efficients (WTC2) is used for proposed fault location method. The next
method. step is to define whether the fault is in the first or second half of the
In [18,19], the traveling wave based fault location principles for transmission line. For this task polarities of first and second voltage
HVDC transmission lines are analyzed. This method is applicable in the traveling wave fronts can be used. Finally, by using the single ended
single-ended or double-ended methods. The maximum fault location traveling wave theory and using the Bewley-Lattice diagram, the loca­
error is about 0.3% of the length of the transmission line. But the faulty tion of the fault is estimated. In the following, Clarke and karrenbauer
half of transmission line identification method is not presented in this transformations for modal components extraction, DWT and traveling
paper. In [8], a single-ended traveling wave based fault location method wave based fault location principle based on single ended data are
for segmented HVDC transmission line (an overhead line combined with presented.
an underground cable) is presented. In this paper, the normalized en­ In this paper, Clarke’s transformation is used to decoupling the three-
ergies are used as the input to a binary Support Vector Machine (SVM) phase voltage in the HVAC system. The Clarke’s transform is formulated
classifier for faulty section identification (underground cable or over­ as follows:
head line). In [20], by using the single-end voltage signal processing and ⎡ ⎤
1 1
mathematical morphology method, the reverse of the voltage traveling 1 − −
⎢ 2 2 ⎥⎡ ⎤
waves (RVTW) is calculated and used for fault location in HVDC trans­ ⎡ ⎤
Vα
⎢
⎢ √̅̅̅ √̅̅̅ ⎥
⎥ Va
mission lines. The accuracy of fault location in this method is about 98% 2 ⎢ 3 3 ⎥⎣ V ⎦
⎣ Vβ ⎦ = ⎢ 0 − ⎥ b (1)
and the measurement error is about 1.44%. In [21], wavelet-based V0
3⎢
⎢ 2 2 ⎥ ⎥ Vc
⎣1 1 1 ⎦
electromagnetic time reversal method using the current signals of both
terminals of the transmission line is used to locate faults in VSC-HVDC 2 2 2
transmission lines. In [7,22], a fault traveling wave-based location
method for hybrid AC lines (underground cable and overhead trans­ Where Va, Vb and Vc represent the voltage values of the phase “A”, “B”
mission line) is proposed. In [7], the extreme-point symmetric mode and “C” respectively; and Vα, Vβ and V0 represent the modal values. The
decomposition (ESMD) method and Teager energy operator (TEO) are first mode (1-mode) and second mode (2-mode) are known as the aerial
used to detect the arrival time of traveling waves. mode and the 0-mode is referred to as the ground mode [12]. In this
In the papers presented above, fault location in HVAC or HVDC paper, 1-mode (Vα) is used to locate all AC faults, except for BC fault,
transmission lines has been investigated using the traveling wave 2-mode (Vβ) is used.
method. In some papers, studies have been conducted on electromag­ The bipolar coupled line voltages can be decoupled to earth-mode (0-
netic coupling between AC and DC lines in HPOTLs, and in a few pub­ mode) and line-mode (1-mode) components according to the decoupling
lished papers, fault detection in HPOTLs has been investigated [23–31]. karrenbauer phase-mode matrix that given as follows [32]:

2
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

Where, ψ(0) is the mother wavelet.a = am m

0 and b = nb0 a0 are repre­
sented as scaling and the translation parameter, respectively [34].
Some of the common wavelets used in the analysis of the transient
studies of the power system, Signal processing and etc., are for example,
Meyer, Daubechies (db), Morlet, Mexican Hat wavelet and etc. Modal
voltage and current signals are processed using DWT to obtain the first
level detail coefficients (D1). Haar wavelet, written as db1 is the
simplest wavelet imaginable and is used in this paper as the mother
Fig. 1. Decomposition tree of discrete wavelet transform (DWT) (Four level of wavelet in the analysis of HVDC waveforms. Also, db4 is used as the
decomposition). mother wavelet in the analysis of HVAC waveforms.

[ ]
1 1
[ ][ ] 2.2. Traveling wave based fault location principle
u1 − 1 u+
= √̅̅̅ (2)
u0 2 1 1 u−
In this paper, the single-ended traveling waves fault location method
Where u1 is 1-mode voltage and u0 is 0-mode voltage. Since um is in is used. As we know, fault location methods based on the single-ended
mode form, positive and negative voltage are decoupled and the effect of transmission line signals have higher reliability and speed due to the
mutual inductances is eliminated. Considering that the 1-mode param­ lack of dependence on data from the remote end of the transmission line
eters are more stable than 0-mode parameters, the 1-mode parameters and also the elimination of the communication channel between the two
are used for fault location calculations in the HVDC line in this paper ends of the transmission line. In single-ended fault location methods, all
[32]. measurements are performed simultaneously on one end of the trans­
mission line. In the fault location methods based on traveling waves, the
2.1. Discrete wavelet transform (DWT) location of fault is calculated based on the reflection time of the trav­
eling waves from the fault point to the measuring terminal. In grounded
Wavelet analysis is a very powerful and practical signal processing faults, in addition to the waves received at the measuring terminal from
method. This method is suitable for detecting sudden changes in a signal the fault point, some waves are also reflected from the remote end of
(transient phenomena). Therefore, it is widely used in fault location transmission line towards the measuring terminal. In the traveling
algorithms in electrical power systems. It should be noted that wavelet wave’s theory and using the Bewley-Lattice diagram, the waves received
analysis does not use a time-frequency, but rather a time-scale region at the measuring terminal are similar to each other for faults with a
[33]. The analyzed signal is decomposed into different scales using a symmetrical position relative to the middle of the transmission line.
wavelet analysis function called the ‘mother wavelet’. This wavelet is Therefore, the fault position in the first or second half of the trans­
scaled and translated to match an input signal locally. The subsequent mission line should be determined using special algorithms [12]. Fig. 2
calculated wavelet coefficients represent the correlation between the (a) and (b) shows waves propagation based on Bewley-Lattice diagrams
(scaled) wavelet and the signal. Discrete Wavelet Transform (DWT) is for grounded and ungrounded faults, respectively. In the Bewley-lattice
introduced as a set of low-pass (called by H) and high-pass (called by G) diagram, the following properties exist:
filters. Such a structure, with four level of decomposition, has been
shown in Fig. 1, where the original signal is decomposed into the • In the path of traveling waves, time is increasing.
approximation (aj) and detail (dj) coefficients. The discrete wavelet • The position of any wave at any time can be deduced directly from
transform of a discrete signal x(k) is defined as: the diagram.
( ) • The total voltage/current at any point, at any instant of time is the
1 ∑ k − nb0 am0 superposition of all the waves which have arrived at that point up
DWT(m, k) = √̅̅̅̅̅
m
x(n)ψ m
(3)
a0 n a0

Fig. 2. Waves propagation based on Bewley-Lattice diagrams for (a) grounded and (b) ungrounded faults.

3
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

faults of the first half and the second half of the transmission line,
respectively.
ν.(Ts2 − Ts1 )
Xf = (5)
2

ν.(Ts2 − Ts1 )
Xf = L − (6)
2

Where Xf is the distance of the fault from the terminal S, ν is the wave
propagation velocity in the used mode of voltage and Ts1 and Ts2 are the
peaks of the first and second waves on the DWT coefficients received to
the terminal S, respectively.

Fig. 3. Small element of a transmission line [36].

2.3. Faulty half of transmission line identification

until that instant of time, displaced in position from each other by The faulty half identification method presented in [35,36] is used in
intervals equal to the difference in their time of arrival. this paper. According to the of traveling waves theory, when a short
• The history of the wave is easily traced. It is possible to find where it circuit (SC) fault occurs in the transmission line, the voltage traveling
came from and just what other waves went into its composition. waves with the same sign move from the fault point to both end of the
transmission line (For example, in Fig. 4, the voltage traveling waves are
In this Fig. 2, Xf indicates the fault distance from the measuring shown with a positive sign). When the traveling wave reaches the end of
terminal (terminal S) and L is the length of the transmission line. the transmission line, some energy passes through the termination cir­
According to the wave propagation model shown in Fig. 2, the cuit and the rest is reflected towards the fault point. The reflection co­
location of the fault can be estimated based on the time of the waves efficient and its sign at the station termination depend on the
received at terminal S. Following is the formulation of the problem for termination circuit, but the sign of the reflection coefficient of the
ungrounded and grounded faults: voltage is opposite to the current. If the station termination is capacitive,
the voltage reflection coefficient is negative and if the station termina­
2.2.1. Ungrounded faults tion is inductive, the voltage reflection coefficient is positive. But in both
As mentioned in literatures, in ungrounded faults such as phase-to- cases, the transmission coefficient is positive. The return traveling wave
phase fault in an AC transmission lines or pole-to-pole in a DC trans­ from the end of the transmission line hits the fault point (SC) and some
mission lines, there is no significant reflection from remote end terminal of its energy passes through the fault and the rest of it is reflected to­
during the fault transient, and the traveling waves between the fault wards the end of the transmission line and this movement continues
point and transmission line terminals are moved (Fig. 2 (b)). Therefore, until the energy of the traveling wave becomes zero. Since the faults
by measuring the time delay between two consecutive peaks in the investigated in this paper are SC fault types, the reflection coefficient is
discrete wavelet transform coefficients of the fault signal recorded in the positive. It should be noted that the transmission coefficient of traveling
desired scale and taking the product of the wave velocity and half of this waves is always positive at the fault point. In the following, the traveling
time delay, the distance to the fault can easily be calculated for these wave’s equations and the reflection and transmission coefficients in
kinds of faults [12,35]. In the following, the equation of calculating the transmission lines are presented.
fault location for ungrounded faults is presented. In the studies presented in this section and to show the traveling
ν.(T
′ ′
− T s1 ) waves equations, the line is assumed loss-less. Fig. 3 shows a small
(4) element of a transmission line with an inductance L (H/m), a capaci­
s2
Xf =
2
tance C (F/m) and an elementary length Δx will have inductance LΔx
Where Xf is the distance of the fault from the terminal S, ν is the wave and capacitance CΔx and as shown [36].
propagation velocity in the used mode of voltage and Ts1 and Ts2 are the
′ ′
Voltage across the element and the current to charge the elementary
peaks of the first and second waves on the DWT coefficients received to capacitance CΔx are defined as follows:
the terminal S, respectively. ∂V ∂I
=− L (7)
∂x ∂t
2.2.2. Grounded faults
When a grounded fault occurs in the transmission line, unlike ∂I ∂V
=− C (8)
ungrounded faults, in addition to the traveling waves received at the ∂x ∂t
terminal S from the fault point, waves are also reflected from the remote By differentiating Eq. 7 with respect to ‘x’ and Eq. 8 with respect to ‘t’
end terminal towards the terminal S Fig. 2 (a)). Also, depending on the and integrating them together, equations for the voltage and current of
location of the fault, the waves reflected from the remote end terminal the transmission line can be obtained separately:
(terminal R) may be received at the terminal S before or after the second
wave reached the terminal S from the fault point. By using the Bewley- ∂2 V ∂2 I ∂2 I ∂2 V
=− L , =− C 2
Lattice diagram, it can be concluded that if the fault occurs in the first ∂x 2 ∂x∂t ∂x∂t ∂t
half of the transmission line, the first and second waves received at
terminal S are reflected from the fault point. But if the fault occurs in the ∂2 V ∂2 V
= LC 2 (9)
second half of the transmission line, the first wave from fault point and ∂x 2
∂t
the second wave have been reflected from the remote end terminal and
reached the terminal S [12,35]. According to the previous explanations ∂2 I ∂2 I
= LC 2 (10)
and assuming that the fault detection and classification algorithm has ∂x2 ∂t
detected the type of fault as grounded fault, the faulty half of the By solving Eq. 9, it can be seen that the current equation will be in the
transmission line should be identified. After faulty half determination, form of Eq. 11:
by using Eq. (5) and ((6), the fault location can be estimated for the

4
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

Fig. 4. Bewley-Lattice diagram for voltage traveling waves for (a) first and (b) second half transmission line faults.

[ ]
t
I = f x ± √̅̅̅̅̅̅ (11) I3 =
V3
(16)
LC ZB
As we know, (L/C)− 1/2 is a velocity. Therefore, using Eq. 11, the According to the explanations given earlier, since the traveling
current equation is obtained as Eq. 12. Also, Eqs. 7 and 12 have been waves of voltage and current in the direction of minus x have opposite
used to obtain the Eq. 13 for voltage: signs, the negative sign is included in Eq. 15. If voltage and current are to
I(x, t) = f1 (x − νt) + f2 (x + νt) be continuous at the junction, it follows that:

∂V ∂I ′ ′
(12) V1 + V2 = V3 , I1 + I2 = I3 (17)
= − L = Lν[f 1 (x − νt) − f 2 (x + νt)]
∂x ∂t Using Eqs. (14–17) and combining them, the reflection and trans­
V(x, t) = Lν[f1 (x − νt) − f2 (x + νt)] = Z0 f1 (x − νt) − Z0 f2 (x + νt) (13) mission coefficients for the voltage traveling wave can be obtained as
follow:
In the voltage and current equations presented above, f1(x-νt) and ( )
f2(x+νt) represent traveling waves moving in the direction of plus x and V2 =
ZB − ZA
V1 →V2 = aV1 , − 1 ≤ a ≤ 1 (18)
moving in the direction of minus x with a velocity ν, respectively. In ZB + ZA
other words, traveling waves in the forward and backward directions. ( )
2ZB
The difference between voltage and current is characteristic impedance V3 = V1 →V3 = bV1 , 0 ≤ b ≤ 2 (19)
(Z0). Since the signs of the terms in the voltage and current equations ZB + ZA
show the direction of the traveling waves in the transmission line, the √̅̅̅̅
sign of the second term in the voltage equation is very important. We a=
ZB − ZA
, b=
2ZB
, ZA =
L
(20)
also note that current and voltage waves traveling in the positive di­ ZB + ZA ZB + ZA C
rection of x have the same sign, whereas those traveling the negative In the above equations, ‘a’ is the reflection coefficient and ‘b’ is the
direction have opposite sign [36]. In the following, the reflection and transmission coefficient. Examples of discontinuity points are short
transmission coefficients of traveling waves in transmission lines have circuit points and transmission line terminals (both end of transmission
been investigated. line), for which reflection and transmission coefficients should be
When the wave reaches a discontinuity in the transmission line so calculated. In short circuit faults points, the fault impedance (ZB) in
that the characteristic impedance of the transmission line changes, some normal faults is lower than the characteristic impedance (ZA) of the line,
of the wave is reflected back and the rest of the wave penetrates beyond and according to Eqs. 18 and 19, it can be seen that for the voltage
the discontinuity. Consider the junction between lines of characteristic traveling wave, the reflection coefficient will be a negative value and the
impedance ZA and ZB and it is assumed that ZA > ZB [36]. Suppose the transmission coefficient will be a positive value. Therefore, when the
voltage surge reaches the discontinuity with a step function and with voltage traveling wave meets the short circuit point, the reflected
amplitude V1. At this moment, the amplitude of the corresponding voltage wave will have the opposite sign to the original wave. For the
current wave is equal to: open circuit (OC) faults, if we assume the impedance of the open circuit
V1 to be infinite (ZB →∞), the sign of the reflection coefficient and trans­
I1 = (14) mission coefficient for the voltage traveling wave will be the same as the
ZA
initial wave. But, the calculation of the reflection and transmission co­
If the reflected and transmitted traveling wave’s voltage is equal to efficients at both ends of the transmission line depends on the termi­
V2 and V3, respectively, the reflected and transmitted traveling wave’s nation circuit, and different values are obtained depending on whether
current are equal to: the circuit is capacitive or inductive. Since the termination circuit of the
− V2 presented transmission line model in this paper is inductive, in this
I2 = (15) section the reflection and transmission coefficients for inductive mode
ZA
are presented.

5
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

Fig. 5. Method of voltage signals sampling (create a data window for fault location).

Using Eqs. 18, 19 and 21, the reflection and transmission coefficients
can be calculated in terms of ’S’ as follows, assuming that the initial
voltage traveling wave is a step function with amplitude V1:
[ ] [ ]
V1 L1 s − ZA V1 s − (ZA /L1 )
u2 (s) = au1 (s) = = (22)
s L1 s + ZA s s + (ZA /L1 )
[ ] [ ]
V1 2L1 s V1 2s
u3 (s) = bu1 (s) = = (23)
s L1 s + ZA s s + (ZA /L1 )

Table 1
HPOTL system specifications.
HVAC SYSTEM HVDC SYSTEM

Power 400 MW Power 1000

MW
Voltage 400 kV Voltage ± 500
kV
Frequency 50 Hz Current (per pole) 1000 A
AC Trans. MVA 750 MVA Vac-Rec (L-L)/ Vac- 345/230
Fig. 6. Flowchart of the proposed fault location algorithm. Inv (L-L) kV
AC Filter 11th, 13th, HP and fixed DC Filter 12th, HP
shunt capacitor
Suppose the voltage wave travels on the transmission line with TRANSMISSION LINE CONDUCTORS
characteristic impedance ZA and impinges the terminal equipment with Total Bundled 3 SAG (all 10 m
ZB impedance. According to Eq. 21, ZA is independent of ‘S’ (frequency), conductors)
Sub-Conductors 0.4 m Ground Resistivity 100 Ω.m
but the value of ZB (inductive impedance) in depends on ‘S’. Spacing
√̅̅̅̅̅̅ √̅̅̅̅ DC Resistance 0.0394 Ω/km Number of Ground 2
Ls L Wire
ZA = = , ZB = L1 s (21)
Cs C

Fig. 7. (a) HPOTL model and (b) HPOTL tower configuration.

6
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

Table 2
Types of faults in the HPOTLs and the components used for fault location.
Faults classification Fault types Component used for fault location

DC faults Pole to ground SPGF Second detail component of line mode (1-mode) of DC voltage
Pole to pole PPF (Vdc1-Scale 2-Haar wavelet)
Pole to pole to ground PPGF
AC faults Phase to ground SLGF First detail component of areal mode (1-mode) of AC voltage (Vac1-Scale 1-db4 wavelet)
Phase to phase LLF Except for BC fault (phase B to C)
Phase to phase to ground LLGF First detail component of areal mode (2-mode) of AC voltage (Vac2-Scale 1-db4 wavelet)
Three phase LLLF
AC/DC intersystem faults Phase to pole LPF First detail component of areal mode (1-mode) of AC voltage
Phase to pole to ground LPGF (Vac1-Scale 1-db4 wavelet)
Phase to two pole LPPF Except for BCP and BCN fault (phase B to C to pole)
Phase to two pole to ground LPPGF First detail component of areal mode (2-mode) of AC voltage
Phase to phase to pole LLPF (Vac2-Scale 1-db4 wavelet)
Phase to phase to pole to ground LLPGF and
Two phase to two pole LLPPF Second detail component of line mode (1-mode) of DC voltage
Two phase to two pole to ground LLPPGF (Vdc1-Scale 2-Haar wavelet)
Three phase to pole LLLPF
Three phase to pole to ground LLLPGF
Three phase to two pole LLLPPF
Three phase to two pole to ground LLLPPGF

[ ]
1 α According to the sign of voltage traveling waves received at the
(L1 / ZA ) = (1 / α) → u2 (s) = V1 − (24)
(s + α) s(s + α) terminal ‘S’, it can be found that the fault occurred in the first or in the
second half of the transmission line. As can be seen in Fig. 4, when the
[ ]
2 fault occurs in the first half of the transmission line, the first and second
u3 (s) = V1 (25)
(s + α) waves received at terminal ‘S’ have opposite signs (+/- or -/+), and
when the fault occurs in the second half of the transmission line, the first
The reflection and transmission coefficients in terms of ‘t’ are
and second waves received at terminal ‘S’ have the same (+/+ or -/-).
calculated as follows:
This issue can be understood by comparing polarity of the first and
V2 (t) = − V1 (1 − 2e− αt ) (26) second observed voltage traveling waves. In this paper, the 1-mode of
AC voltage is used to faulty half of transmission line identification in AC
V3 (t) = 2V1 e− αt
(27) and AC/DC intersystem faults, and the 1-mode of DC voltage is used to
faulty half of transmission line identification in DC faults.
Using Eqs. 26 and 27, it can be concluded that when the wave arrives
at its terminals, momentarily it appears like an open circuit and so that
ultimately it appears as a short circuit [36]. Therefore, the voltage wave 2.4. Proposed fault location method
is reflected and transmitted at the moment of impact with the same sign
as the initial wave, and the sign of the reflection and transmission co­ The presented method for fault location is that the AC and DC voltage
efficients becomes positive. According to the explanations provided in signals from sending end of transmission lines are sampled with a
this section, Bewley-Lattice diagram for voltage traveling waves for first sampling rate of 100 kHz and stored in a moving data window with a
and second half transmission line faults is presented in Fig. 4. length of 20 ms. As soon as the fault occurs and the fault is detected by

Fig. 8. (Up) 1-mode of AC voltage and (down) AC voltage WTC2 at scale-1 under SLGF (A to ground fault) at x=318 km.

7
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

Fig. 9. AC voltage WTC2 at scale-1 under LLF (A to B fault) at x=18 km.

Fig. 10. (Up) 1-mode of DC voltage and (down) DC voltage WTC2 at scale-2 under SPGF (positive pole to ground fault) at x=111 km.

the protection algorithms, the fault detection signal is sent to fault As mentioned above, the method of voltage signals sampling (create
location algorithm and the sampling of AC and DC voltage will continue a data window for fault location) and the flowchart of the proposed fault
for 10 ms and is stored in the data window. There are now 10 ms of pre- location algorithm are shown in Figs. 5 and 6, respectively.
fault voltage samples and 10 ms of post-fault voltage samples in the data
window, which is enough to locate the fault. By using signal processing 3. Hybrid parallel HVAC/HVDC overhead transmission line
based on DWT on the data window, the location of the fault can be (HPOTL) modeling
found. In the following, the steps of the proposed method for fault
location in HPOTL lines are presented. In this section, the used model of hybrid parallel HVAC/HVDC
The process of the proposed fault location method for HPOTLs goes overhead transmission line is presented to study and simulate fault
through the following steps: location in AC and DC lines. Fig. 7 shows the model of a HPOTL on the
same tower and tower configuration simulated in this paper. Electro­
- Sampling of AC and DC voltage signals from sending end of trans­ magnetic coupling between AC and DC lines is considered as one of the
mission lines with a sampling rate of 100 kHz and stored in a moving main challenges in HPOTLs. Therefore, voltage and current components
data window with a fixed length of 20 ms. are induced in HVAC and HVDC transmission line under different
- When the fault detection signal is received by the fault location al­ operating conditions, normal and transient conditions (fundamental
gorithm, the sampling of AC and DC voltage will continue for 10 ms frequency components in the DC line and the DC component in the AC
and is stored in the data window. line). Voltage and current equations of electromagnetic coupling be­
- Perform modal transformations (Clarke and Karrenbauer trans­ tween lines in parallel transmission lines are presented in the appendix.
formations) on AC and DC voltage signals on data window. It should be noted that, transmission lines transposition is considered in
- Apply DWT to voltage signal. this model and all voltage and current waveforms shown are from the
- Calculation of squared wavelet transform coefficients (WTC2) for sending end of transmission line. Simulated model in PSCAD software
voltage signals. includes a 400 kV, 400 MW, 50 Hz ac single circuit with a ±500 kV,
- Determination of faulty half of transmission line for grounded faults 1000 MW dc circuit connected into the same receiving and sending end
by comparing polarity of the first and second observed voltage ac systems. DC section is a bipolar CIGRE 12 pulse HVDC model and the
traveling waves. DC controls are based on the CIGRE benchmark model [37]. The total
- Using traveling wave theory and Bewley-Lattice diagram to locate length of HPOTL is 400 km. As shown in Fig. 7, AC filters in bus A and
the fault. bus B (11th, 13th and High Pass filter) are added to absorb the harmonics
generated by the converter as well as together with a fixed shunt

8
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

Fig. 11. (Up) 1-mode of AC voltage, (middle) AC voltage WTC2 at scale-1 and (down) DC voltage WTC2 at scale-2 under LPPGF (phase A to PN to ground fault) at
x=218 km.

Table 3
Some examples for AC fault location in AC transmission line in HPOTLs.
Fault type Fault Xact 1st sample 2nd sample ΔS Polarity of the 1st and 2nd wave Faulty half Xcal ABS (error %)
(km) sample (km)

AC Faults SLGF AG 24 9 25 16 Opposite First 23.60 0.10

98 35 101 66 Opposite First 97.35 0.16
215 73 197 124 Same Second 217.10 0.52
318 106 164 58 Same Second 314.45 0.89
LLF AB 18 6 18 12 - - 17.70 0.07
131 45 132 87 - - 128.33 0.67
218 74 221 147 - - 216.83 0.29
341 115 345 230 - - 339.25 0.44
BC 41 15 42 27 - - 39.83 0.29
117 38 117 79 - - 116.53 0.12
205 68 206 138 - - 203.55 0.36
362 121 367 246 - - 362.85 0.21
LLGF ACG 19 6 19 13 Opposite First 19.18 0.04
122 43 125 82 Opposite First 120.95 0.26
221 75 195 120 Same Second 223.00 0.50
346 117 153 36 Same Second 346.90 0.22
LLLF ABC 31 10 31 21 - - 30.98 0.01
109 36 111 75 - - 110.63 0.41
231 77 232 155 - - 228.63 0.59
333 111 335 224 - - 330.40 0.65

capacitor to supply reactive power to the converter. DC smoothing re­ lines) are parallel to each other in the entire line and are on the same
actors installed in series with the poles of the HVDC transmission line are tower (Fig. 7). The frequency dependent line model is used for transient
0.5968 H. Also a high-pass DC filter tuned to the 12th harmonic is placed studies. The simulation time step in 10 μs and the sampling rate is 100
on the DC side. HPOTL system specifications are given in Table 1. kHz. Simulations are carried out by using PSCAD software and MATLAB
software (MATLAB Wavelet Toolbox) is used for data analysis and signal
4. Simulations results processing. The voltage traveling wave velocity for areal mode is
assumed to be 2.95 × 105 km/s. As mentioned before, in this paper, it is
Simulation results for fault location in HPOTLs according to the assumed that the occurrence of the fault (AC, DC and AC/DC intersystem
model presented in Section 3 are presented in this section. The length of fault) and its classification (detection of the faulty pole and/or faulty
the HPOTL is 400 km, and both systems (HVAC and HVDC transmission phase) has been detected by a protection algorithm. Since the presented

9
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

Table 4
Some examples for DC fault location in DC transmission line in HPOTLs.
Fault type Fault Xact 1st sample 2nd sample ΔS Polarity of the 1st and 2nd wave Faulty half Xcal ABS (error %)
(km) sample (km)

DC Faults SPGF PG 21 5 21 16 Opposite First 23.60 0.65

111 37 113 76 Opposite First 112.10 0.28
234 77 189 112 Same Second 234.80 0.20
371 125 144 19 Same Second 371.98 0.24
NG 32 9 33 24 Opposite First 35.40 0.85
127 41 129 88 Opposite First 129.80 0.70
254 85 185 100 Same Second 252.50 0.38
346 117 153 36 Same Second 346.90 0.22
PPF PN 18 5 17 12 - - 17.70 0.07
105 33 105 72 - - 106.20 0.30
243 81 245 164 - - 241.90 0.27
351 117 353 236 - - 348.10 0.72
PPGF PNG 15 5 17 12 Opposite First 17.70 0.68
118 41 121 80 Opposite First 118.00 0.00
210 69 200 131 Same Second 206.78 0.81
348 117 153 36 Same Second 346.90 0.28

fault location method is based on the singe end data of the transmission 4.2. Case 2: AC fault: phase to phase (A to B) fault (LLF)
line, only the voltage data of the sending end of the transmission line has
been used to fault location and faulty half determination. The second case is the phase to phase (A to B) fault (LLF) in the AC
Different type of balanced and unbalanced faults such as AC, DC and line at 18. Since the fault under study is an ungrounded fault, it is not
AC/DC intersystem faults at different locations along the line are required to identify the faulty half of the transmission line and fault
simulated. The types of faults in the HPOTLs and the components used location is calculated using traveling wave theory for ungrounded faults
for fault location are presented in Table 2. As shown in this Table, in (Eq. 4). The waveforms of the WTC2 for AC voltage at scale-1 for LLF are
HPOTLs there are 5 types of AC faults, 3 types of DC faults and 12 types shown in Fig. 9.
of AC/DC intersystem faults. All of AC, DC and intersystem faults have ′ ′
ν.(T − T s1 ) 2.95 × 105 × (0.00018 − 0.00006)
been simulated for fault location method validation. But in this section, Xf =
s2
= = 17.7 km
fault location is presented only for 4 different cases (two case for AC 2 2
faults and one case for each of DC and AC/DC intersystem faults) using Using Eq. 28, the fault location error for this case is calculated 0.07%.
the proposed method based on traveling waves. Also, the effectiveness of
the proposed fault location method is evaluated using the error which is 4.3. Case 3: DC fault: positive pole to ground fault (SPGF)
calculated as a percentage of the total system length as:
⃒ ⃒
⃒Actual fault location − calculated fault location⃒ The third case is the positive pole to the ground fault (SPGF) in the
Error (%) = ⃒⃒ ⃒ × 100 (28)
Total length of transmission line ⃒ DC line at 111 km. The waveforms related to 1-mode DC voltage from
the sending end of the transmission line and the WTC2 for DC voltage at
scale-2 are shown in Fig. 10.
4.1. Case 1: AC fault: phase A to ground fault (SLGF)
As can be seen in Fig. 10 (up), the polarity of the first and second
voltage traveling waves are opposite. According to what was presented
The first case is the phase A to ground fault (SLGF) in the AC line at
at the end of section 2.3, it can be concluded that the fault occurred in
318 km from the sending end of the transmission line. Since the fault
the first half of the transmission line. In Fig. 10 (down), the first and
under study is a grounded fault, the faulty half of the transmission line
second DC voltage traveling waves received at sending end of trans­
should be identified. The waveforms of 1-mode AC voltage from the
mission line and their arrival time are specified. Therefore, by using the
sending end of the transmission line and the WTC2 for AC voltage at
time difference of two received waves and using Eq. 5, the location of the
scale-1 for SLGF are shown in Fig. 8.
fault can be obtained as follows:
As can be seen in Fig. 8 (up), the polarity of the first and second
voltage traveling waves are the same (both are negative). According to ν.(Ts2 − Ts1 ) 2.95 × 105 × (0.00112 − 0.00036)
what was presented at the end of section 2.3, it can be concluded that the Xf =
2
=
2
= 112.1 km
fault occurred in the second half of the transmission line. In Fig. 8
Using Eq. 28, the fault location error for this case is calculated
(down), the first and second DC voltage traveling waves received at
0.275%.
sending end of transmission line and their arrival time are specified.
Therefore, by using the time difference of two received waves and using
Eq. 6, the location of the fault can be obtained as follows: 4.4. Case 4: AC/DC intersystem fault- phase to two pole (B to PN) to
ground fault (LPPGF)
ν.(Ts2 − Ts1 ) 2.95 × 105 × (0.00164 − 0.00106)
Xf = L − = 400 −
2 2 The fourth case is the phase to two pole (B to PN) to ground fault
= 314.45 km (LPPGF) between the AC and DC line at 218 km from the sending end of
the transmission line. Similar to case 1 and 3, in this case also the faulty
Using Eq. 28, it is possible to estimate the error of fault location at
half of the transmission line should be identified. As shown in Table 2,
proposed method. The fault location error in this case is calculated as
the number of intersystem faults is large, and in this section, the LPPGF
follows:
⃒ ⃒ is examined as an example of this type of fault. Since the AC/CD inter­
⃒318 − 314.45⃒ system faults occurs between the phases of the AC system and the poles
Error (%) = ⃒⃒ ⃒ × 100 = 0.8875%
400 ⃒ of the DC system and involves both systems in fault, in this paper in
order to locate these faults, the AC and DC voltage components are used.
AC voltage component is used to faulty half of transmission line iden­
tification in intersystem faults type. But the proposed fault location

10
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

Table 5
Some examples for AC/DC intersystem fault location in HPOTLs.
Fault type Fault Xact 1st 2nd ΔS Polarity of the 1st Faulty Xcal Error Final ABS (total
(km) sample sample sample and 2nd wave half (km) % Xcal error %)
(km)

AC/DC LPF AP 35 13 37 24 - - 35.40 0.10 36.87 0.47

intersystem 13 39 26 - - 38.35 0.84
Faults 350 117 353 236 - - 348.10 -0.47 349.57 0.11
117 355 238 - - 351.05 0.26
LPGF BPG 190 65 193 128 Opposite First 188.80 -0.30 189.53 0.12
62 191 129 Opposite First 190.28 0.07
371 125 144 19 Same Second 371.98 0.24 371.23 0.06
125 145 20 Same Second 370.50 -0.13
LPPF APN 122 41 125 84 - - 123.90 0.48 122.42 0.11
43 125 82 - - 120.95 -0.26
221 73 225 152 - - 224.20 0.80 221.25 0.06
75 223 148 - - 218.30 -0.67
LPPGF BPNG 45 13 45 32 Opposite First 47.20 0.55 45.72 0.18
15 45 30 Opposite First 44.25 -0.19
318 105 161 56 Same Second 317.40 -0.15 317.4 0.15
106 162 56 Same Second 317.40 -0.15
LLPF BCP 130 45 133 88 - - 129.80 -0.05 130.53 0.13
42 131 89 - - 131.28 0.32
230 77 233 156 - - 230.10 0.03 228.62 0.34
77 231 154 - - 227.15 -0.71
ABP 62 25 69 44 - - 64.90 0.73 63.42 0.36
23 65 42 - - 61.95 -0.01
341 113 345 232 - - 342.20 0.30 339.98 0.25
115 344 229 - - 337.78 -0.81
LLPGF ABNG 29 9 29 20 Opposite First 29.50 0.13 28.76 0.06
11 30 19 Opposite First 28.03 -0.24
381 138 153 15 Same Second 377.88 -0.78 378.61 0.60
126 140 14 Same Second 379.35 -0.41
LLPPF ACPN 110 41 117 76 - - 112.10 0.53 110.62 0.16
39 113 74 - - 109.15 -0.21
242 81 245 164 - - 241.90 -0.02 241.9 0.02
82 246 164 - - 241.90 -0.02
LLPPGF ABPNG 111 37 113 76 Opposite First 112.10 0.28 110.62 0.09
37 111 74 Opposite First 109.15 -0.46
345 113 149 36 Same Second 346.90 0.47 345.42 0.11
115 153 38 Same Second 343.95 -0.26
LLLPF ABCP 25 9 25 16 - - 23.60 -0.35 23.6 0.35
8 24 16 - - 23.60 -0.35
228 77 233 156 - - 230.10 0.53 227.88 0.03
77 230 153 - - 225.68 -0.58
LLLPGF ABCNG 150 49 153 104 Opposite First 153.40 0.85 150.45 0.11
51 151 100 Opposite First 147.50 -0.63
355 117 149 32 Same Second 352.80 -0.55 353.53 0.37
118 149 31 Same Second 354.28 -0.18
LLLPPF ABCPN 61 21 61 40 - - 59.00 -0.50 60.47 0.13
20 62 42 - - 61.95 0.24
320 105 321 216 - - 318.60 -0.35 320.07 0.02
107 325 218 - - 321.55 0.39
LLLPPGF ABCPNG 177 57 177 120 Opposite First 177.00 0.00 176.26 0.18
59 178 119 Opposite First 175.53 -0.37
210 69 197 128 Same Second 211.20 0.30 211.2 0.30
71 199 128 Same Second 211.20 0.30

method in this paper suggests a new method to increase the accuracy of voltage at scale-2 for LPPGF are shown in Fig. 11.
calculating the fault location for AC/DC intersystem faults. In this paper, As can be seen in Fig. 11 (up), the polarity of the first and second
for intersystem fault location, the average location estimated by AC voltage traveling waves are same. According to what was presented at
voltage and DC voltage (which was used separately to locate the AC and the end of section 2.3, it can be concluded that the fault occurred in the
DC transmission line) is used. The improvement of fault location accu­ second half of the transmission line. Using Fig. 11 (middle) and 11
racy in this method is shown in Table 9. The waveforms of 1-mode AC (down), the fault location can be obtained in two ways using WTC2 of AC
voltage from the sending end of the transmission line and the squared voltage at scale-1 and WTC2 of DC voltage at scale-2, which is presented
wavelet transform coefficient (WTC2) for AC voltage at scale-1 and DC below.

ν.(Ts2 − Ts1 ) 2.95 × 105 × (0.00198 − 0.00073)

UsingACvoltagecomponent : Xf = L − = 400 − = 215.62 km
2 2
ν.(Ts2 − Ts1 ) 2.95 × 105 × (0.00196 − 0.00072)
UsingDCvoltagecomponent : Xf = L − = 400 − = 217.1 km
2 2

11
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

Table 6
Performance of the fault location method under different fault resistance.
Fault type Fault Xact Rf 1st 2nd ΔS Polarity of the Faulty Xcal Error Final ABS (total
(km) (Ω) sample sample sample 1st and 2nd wave half (km) % Xcal error %)
(km)

DC Faults SPGF PG 100 100 33 101 68 Opposite First 100.30 0.08 100.30 0.08
100 200 33 101 68 Opposite First 100.30 0.08 100.30 0.08
100 500 33 101 68 Opposite First 100.30 0.08 100.30 0.08
PPGF PNG 312 100 105 165 60 Same Second 311.50 -0.13 311.50 0.13
312 200 105 165 60 Same Second 311.50 -0.13 311.50 0.13
312 500 105 165 60 Same Second 311.50 -0.13 311.50 0.13
AC Faults SLGF AG 270 100 90 180 90 Same Second 267.25 -0.69 267.25 0.69
270 200 90 180 90 Same Second 267.25 -0.69 267.25 0.69
270 500 90 180 90 Same Second 267.25 -0.69 267.25 0.69
LLGF ABG 142 100 46 143 97 Opposite First 143.08 0.27 143.08 0.27
142 200 46 143 97 Opposite First 143.08 0.27 143.08 0.27
142 500 46 143 97 Opposite First 143.08 0.27 143.08 0.27
AC/DC LPGF APG 125 100 41 125 84 Opposite First 123.90 -0.27 124.64 -0.09
intersystem 43 128 85 Opposite First 125.38 0.09
Faults 125 200 41 125 84 Opposite First 123.90 -0.27 124.64 -0.09
43 128 85 Opposite First 125.38 0.09
125 500 41 125 84 Opposite First 123.90 -0.27 124.64 -0.09
43 128 85 Opposite First 125.38 0.09
LLPGF ABPG 255 100 85 185 100 Same Second 252.50 -0.63 253.97 -0.26
85 183 98 Same Second 255.45 0.11
255 200 85 185 100 Same Second 252.50 -0.63 253.97 -0.26
85 183 98 Same Second 255.45 0.11
255 500 85 185 100 Same Second 252.50 -0.63 253.97 -0.26
85 183 98 Same Second 255.45 0.11

Table 7
Performance of the fault location method under different noisy environment.
Fault type Fault Xact SNR 1st 2nd ΔS Polarity of the Faulty Xcal Error Final ABS (total
(km) sample sample sample 1st and 2nd wave half (km) % Xcal error %)
(km)

DC Faults SPGF PG 152 10 49 153 104 Opposite First 153.40 0.35 153.40 0.35
dB
20 49 153 104 Opposite First 153.40 0.35 153.40 0.35
dB
30 49 153 104 Opposite First 153.40 0.35 153.40 0.35
dB
AC Faults LLF AC 278 10 93 283 190 - - 280.25 0.56 280.25 0.56
dB
20 93 283 190 - - 280.25 0.56 280.25 0.56
dB
30 93 283 190 - - 280.25 0.56 280.25 0.56
dB
AC/DC LPGF ANG 364 10 121 145 24 Same Second 364.60 0.15 363.13 0.22
intersystem dB 123 149 26 Same Second 361.65 -0.59
Faults 20 121 145 24 Same Second 364.60 0.15 363.13 0.22
dB 123 149 26 Same Second 361.65 -0.59
30 121 145 24 Same Second 364.60 0.15 363.13 0.22
dB 123 149 26 Same Second 361.65 -0.59

2nd sample: The time of the second voltage traveling wave received
at the sending end of the transmission line
Using Eq. 28, the fault location error for this case using AC and DC ΔS sample: The time difference between the first and the second
voltage components is equal to 0.595% and 0.225%, respectively. The voltage traveling wave received at the sending
average fault location for this case is 216.36 km and the fault location Final Xcal (km): average fault location calculated by fault location
error is 0.41%. As can be seen, the averaging between the locations method for intersystem faults
estimated by the AC and DC voltage component improves the fault In Table 5, the data of the first row for each fault is obtained using the
location by the proposed method. Tables 3–5 show other examples for WTC2 of DC voltage at scale-2 and the second row using the WTC2 of AC
AC, DC and AC/DC intersystem faults location in HPOTLs, respectively. voltage at scale-1.
The terms used in Tables 3–9 are as follows:
Xact: The actual location of the fault from the sending end of the 5. Sensitivity analysis
transmission line
Xcal: The calculated location of the fault from the sending end of the To sensitivity analysis of the proposed fault location method, the
transmission line by proposed method effects of various factors have been studied as follows.
1st sample: The time of the first voltage traveling wave received at
the sending end of the transmission line (According to the sampling A) Effect of fault resistance: To inspect the sensitivity of the proposed
frequency (100 kHz), each sample is equal to 0.00001 seconds) method to fault resistance, the test system is simulated with a set

12
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

Table 8
Performance of the fault location method under different fault angles.
Fault type Fault Xact θf 1st 2nd ΔS Polarity of the Faulty Xcal Error Final ABS (total
(km) sample sample sample 1st and 2nd wave half (km) % Xcal error %)
(km)

AC Faults SLGF AG 81 0 1677 1731 54 Opposite First 79.65 -0.34 79.65 0.34
30 1837 1891 54 Opposite First 79.65 -0.34 79.65 0.34
45 1927 1981 54 Opposite First 79.65 -0.34 79.65 0.34
60 2007 2061 54 Opposite First 79.65 -0.34 79.65 0.34
90 2177 2231 54 Opposite First 79.65 -0.34 79.65 0.34
AC/DC LPGF AN 280 0 1745 1933 188 - - 277.30 -0.67 278.78 0.31
intersystem 1745 1935 190 - - 280.25 0.06
Faults 30 1905 2093 188 - - 277.30 -0.67 278.78 0.31
1905 2095 190 - - 280.25 0.06
45 1993 2185 192 - - 283.20 0.80 280.99 0.25
1995 2184 189 - - 278.78 -0.31
60 2073 2265 192 - - 283.20 0.80 280.99 0.25
2075 2264 189 - - 278.78 -0.31
90 2245 2433 188 - - 277.30 -0.67 278.04 0.49
2245 2434 189 - - 278.78 -0.31
LLPGF ACPG 332 0 1761 1809 48 Same Second 329.20 -0.70 329.20 0.70
1761 1809 48 Same Second 329.20 -0.70
30 1921 1965 44 Same Second 335.10 0.78 332.15 0.04
1921 1969 48 Same Second 329.20 -0.70
45 2009 2057 48 Same Second 329.20 -0.70 329.20 0.70
2011 2059 48 Same Second 329.20 -0.70
60 2089 2137 48 Same Second 329.20 -0.70 329.20 0.70
2091 2139 48 Same Second 329.20 -0.70
90 2261 2305 44 Same Second 335.10 0.78 332.15 0.04
2261 2309 48 Same Second 329.20 -0.70

Table 9
Comparison of different references in the field of fault location in transmission lines
Author’s name Ref. Publication Transmission line’s model Number of Number of measuring Sampling Mean error
year circuits terminals frequency (%)

L. Xie [7] 2022 AC 1 Double end data NA NA

H. Livani [8] 2014 DC (underground + overhead) 1 Single end data 200 KHz 0.14
H. Livani [9] 2010 AC (underground + overhead) 1 Single end data 100KHz 0.5
A. M. Elhaffar [11] 2008 AC 1 Single and multi-end data 1.25 MHz 0.46
O. Naidu [14] 2018 AC 1 Double end data 1 MHz 0.02
J. Liu [16] 2015 DC 1 Double end data 1 MHz 1
Zheng-You He [17] 2013 DC 1 Single end data 10 MHz 0.17
P. Chen [18] 2006 DC 1 Multi-end data 1 MHz 0.3
P. Chen [19] 2008 DC 1 Multi-end data 1 MHz 0.3
A. Swetha [20] 2011 DC 1 Single end data 80 KHz 1.44
X. Zhang [21] 2019 DC 1 Double end data 50, 100, 200 KHz 0.44
D. Rezaei [22] 2022 AC (underground + overhead) 1 Single end data 1 MHz NA
J. Suonan [32] 2009 DC 1 Double end data 100 KHz 0.32
M. Ando [35] 1985 DC 1 Single end data NA 0.37
Proposed - - Hybrid parallel HVAC/HVDC 2 Single end data 100 KHz 0.33
method overhead

of different fault resistances up to 500 (Ω). Table 6 gives the

Table 10
performance of the fault location method under different fault.
Fault location accuracy in proposed fault location method.
The results presented show that fault resistance has no effect on
Absolute Fault location error Mean Max Max error in the performance of the proposed fault location method.
error (%) error transmission line
(%) (km)
A) Effect of noise: To test the robustness of the proposed fault location
Absolute Fault location error in 0.42 0.85 3.4
method under noisy environment, different white Gaussian
DC faults
Absolute Fault location error in 0.34 0.75 3 noises from 10 dB to 30 dB are added to the measured signals and
AC faults the simulation results are shown in Table 7. It can be inferred that
Absolute Fault location error in 0.41 0.9 3.6 the performance of the proposed fault location method is suitable
AC/DC faults Based on Vdc1- in noisy environments.
Scale 2
Absolute Fault location error in 0.37 0.88 3.52
AC/DC faults Based on Vac1- A) Effect of fault angle: For the fault angle sensitivity analysis, the test
Scale 1 system is simulated with a set of different fault angles on AC
Absolute Fault location 0.25 0.69 2.76 voltage waveform (0, 30, 45, 60 and 90). Table 8 gives the per­
Cumulative error in AC/DC
formance of the fault location method for some faults under
faults Based on
Vdc1-Scale 2 & Vac1-Scale 1
different fault angles. As shown in this table, the proposed
Total mean error in HPOTL 0.33 % method is able to locate the fault with high accuracy in different
fault angles.

13
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

Fig. 12. Differential parameters of mutually coupled parallel transmission lines [38].

6. Discussion and comparison of results AC/DC intersystem faults. As can be seen in Table 9, using both AC and
DC voltage components for intersystem fault location has reduced the
In this section, the obtained results are compared with the results of average and maximum absolute error of fault location. Also, the sensi­
various references that have been published in the field of traveling tivity analysis on the proposed method shows high accuracy, noise im­
wave based fault location in transmission lines. By using indices such as munity, robustness against fault angles and high impedance faults
transmission line’s model, number of measuring terminals, sampling (Table 10).
frequency and mean error of proposed method, it is possible to provide a
suitable comparison between the algorithms presented. In Table 9, the CRediT authorship contribution statement
papers published in this field of research are compared with one another
in different indices presented. Mohammad Fayazi: Conceptualization, Data curation, Formal
analysis, Investigation, Methodology, Software, Validation, Writing –
7. Conclusions original draft. Mahmood Joorabian: Project administration, Supervi­
sion, Validation, Writing – review & editing. Alireza Saffarian: Project
In this paper, a single ended traveling wave based fault location administration, Supervision, Validation, Writing – review & editing.
method is presented for Hybrid Parallel HVAC/HVDC Overhead Trans­ Mehdi Monadi: Methodology, Resources, Writing – review & editing.
mission Lines (HPOTLs) on the same tower using Discrete Wavelet
Transform (DWT). Since the purpose of this paper is to locate the fault in Declaration of Competing Interest
the HPOTLs, assuming that the fault is detected in the HVAC or/and
HVDC transmission line and the faulty lines are identified, the fault The authors declare that they have no known competing financial
location has been done. The classification of faults in the HPOTLs in­ interests or personal relationships that could have appeared to influence
cludes AC, DC and AC/DC intersystem faults. The proposed fault loca­ the work reported in this paper.
tion method in this paper is based on the traveling wave’s theory. For
this purpose, AC and DC transient voltage signals at the sending-end of Data availability
the transmission line are recorded with sampling frequency of 100 kHz,
and then the modal voltage signals are decomposed using the Daube­ No data was used for the research described in the article.
chies4 (Db4-scale 1) wavelet for AC voltage and Haar (Db1-scale 2)
wavelet for DC voltage signals. Finally, the location of fault is calculated
using squared wavelet transform coefficients (WTC2) and Bewley-Lattice Acknowledgements
diagram. Also, for grounded faults, the transmission line faulty-half
identification is done by comparing the polarity of the first and second The authors would like to thank the Shahid Chamran University of
voltage traveling waves observed at the sending end of transmission Ahvaz for financial support under the grant number SCU.EE1401.63 and
line. The results obtained from the simulations show the high accuracy SCU.EE1400.373. Also, the authors would like to thank Khuzestan water
of the proposed method for calculating the fault location. Table 9 shows and power authority (KWPA) and management of the office of applied
the average and maximum absolute error of fault location in AC, DC and and research for support for this research.

Appendix: Electromagnetic coupling in parallel transmission lines

The isolated long transmission line is defined in terms of the series impedance and shunt admittance, both expressed per unit of length. It is
assumed that the line is completely transposed such that the mutual coupling between phases is exactly balanced and a per-phase representation of the

14
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

circuit is permitted [38]. In this case, the voltage and current equations for a single long transmission line in the matrix form is as follows:
[ ] [ ][ ]
VS A B VR A = D = coshγL
= , (29)
IS C D IR C = Yc sinhγL , B = Zc sinhγL

When the two long transmission lines are parallel to each other, the transmission lines equations become more complex because, in addition to the
parameters of the transmission lines, the mutual impedance and the electromagnetic coupling between the lines must be considered. Fig. 12 shows the
differential parameters of two parallel transmission lines that also have electromagnetic coupling. This model is suitable for obtaining voltage and
current equations of parallel transmission lines [38].
According to Fig. 12, the voltage and current equations of A and B lines in the matrix form are as follows:
⎡ ⎤ ⎡ ⎤⎡ ⎤
VAX AAA BAA AAB BAB VAR
⎢ IAX ⎥ ⎢ CAA DAA CAB DAB ⎥⎢ IAR ⎥
⎢ ⎥ ⎢
⎣ VBX ⎦ = ⎣ ABA BBA ABB BBB ⎦⎣ VBR ⎦
⎥⎢ ⎥ (30)
IBX CBA DBA CBB DBB IBR
⏟̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅⏞⏞̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅⏟
M

Eq. (30) is for any x. In various problems, it is usually tended to calculate voltages and currents on the sending end of transmission line. To do this,
we can use the value of L instead of x in the elements of the matrix M in (30). According to Fig. 12, A, B, C and D values in the matrix M in (30) and
detailed voltage and current equations of A and B lines are presented as follow:
dVAX = (zA IAX + zM IBX )dx , dVBx = (zB IBX + zM IAX )dx (31)

dIAX = (yA VAX − yM VBX )dx , dIBX = (yB VBX − yM VAX )dx (32)

Where the two line are coupled by a mutual impedance zM and by a mutual admittance yM. Using Eqs. (31) and (32) and then the derivative of the
result taken with respect to x to obtain the following form for voltage and current of lines.
⎧
⎡ 2 ⎤ ⎪ 2
⎪
d VAx [ ][ ] ⎪ γ AA = zA yA − zM yM
⎪
⎪
⎢ 2 ⎥ 2 2
γ AA − γAB ⎨ γ 2BB = zB yB − zM yM
⎢ dx ⎥ VAX
⎢ 2 ⎥= , (33)
⎣ d VBx ⎦ 2 2
− γ BA γBB V ⎪ 2
BX
⎪ γ AB = zA yM − zM yB
⎪
⎪
⎪
⎩ γ 2BA = zB yM − zM yA
dx2

⎧
⎡ ⎤ ⎪
d2 IAx ⎪
⎪
⎪ γ 2AA = zA yA − zM yM
[ ][ ] ⎪
⎢ 2 ⎥ γ 2AA − γ2AB ⎨ γ2 = zB yB − zM yM
⎢ dx ⎥ IAX
⎢ 2 ⎥ = , BB
(34)
⎣ d IBx ⎦ − γ 2BA γ2BB IBX ⎪ γ 2AB
⎪ = zB yM − zM yA
⎪
⎪
⎪
⎩ γ 2BA
dx2 = zA yM − zM yB

Taking the Laplace transform and using the initial conditions, we can write (33) and (34) as the following matrix equations.
[ ] [ ][ ] [ ][ ]
VAX (s) AAA (s) AAB (s) VAR BAA (s) BAB (s) IAR
= + (35)
VBX (s) ABA (s) ABB (s) VBR BBA (s) BBB (s) IBR
[ ] [ ][ ] [ ][ ]
IAX (s) DAA (s) DAB (s) IAR CAA (s) CAB (s) VAR
= + (36)
IBX (s) DBA (s) DBB (s) IBR CBA (s) CBB (s) VBR

By integrating Eqs. (35) and (36) in the matrix form and transferring to the x domain, (30) is obtained. The matrix elements in Eq. (30) are
calculated as follows.
( 2 ) ( )
γ − γ2BB coshγ1 x − γ22 − γ 2BB coshγ2 x
AAA = 1 2 2
= DAA
γ1 − γ2

γ2AB (coshγ1 x − coshγ2 x)

AAB = = DBA
γ 21 − γ22

γ2BA (coshγ1 x − coshγ2 x)

ABA = = DAB
γ 21 − γ22
( ) ( )
γ 21 − γ2AA coshγ1 x − γ22 − γ 2AA coshγ2 x
ABB = = DBB
γ 21 − γ22
[( ) ] [( ) ]
γ2 γ21 − γ2BB zA − γ2AB zM sinhγ1 x − γ1 γ 22 − γ2BB zA − γ2AB zM coshγ2 x
BAA =
γ1 γ2 (γ21 − γ 22 )
[( ) ] [( ) ]
γ2 γ21 − γ2BB zM − γ2AB zB sinhγ1 x − γ1 γ 22 − γ2BB zM − γ 2AB zB coshγ2 x
BAB =
γ1 γ2 (γ21 − γ 22 )
[( ) ] [( ) ]
γ2 γ21 − γ2AA zM − γ2BA zA sinhγ1 x − γ1 γ 22 − γ2AA zM − γ 2BA zA coshγ2 x
BBA =
γ1 γ2 (γ21 − γ 22 )

15
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

[( ) ] [( ) ]
γ2 γ21 − γ2AA zB − γ2BA zM sinhγ1 x − γ1 γ 22 − γ2AA zB − γ2BA zM coshγ2 x
BBB =
γ1 γ2 (γ21 − γ 22 )
[( ) ] [( ) ]
γ2 γ 21 − γ2BB yA − γ2BA yM sinhγ1 x − γ 1 γ22 − γ 2BB yA − γ2BA yM sinhγ2 x
CAA = 2 2
γ1 γ2 (γ1 − γ 2 )
[( ) ] [( ) ]
γ2 γ 21 − γ2BB yM − γ 2BA yB sinhγ1 x − γ 1 γ22 − γ 2BB yM − γ2BA yB sinhγ2 x
CAB = −
γ1 γ2 (γ21 − γ 22 )
[( ) ] [( ) ]
γ2 γ 21 − γ2AA yM − γ 2AB yA sinhγ1 x − γ 1 γ22 − γ 2AA yM − γ2AB yA sinhγ2 x
CBA = −
γ1 γ2 (γ21 − γ 22 )
[( ) ] [( ) ]
γ2 γ 21 − γ2AA yB − γ2AB yM sinhγ1 x − γ 1 γ22 − γ 2AA yB − γ2AB yM sinhγ2 x
CBB =
γ1 γ2 (γ21 − γ 22 )
( √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ )
1
γ 21 = (zA yA + zB yB − 2zM yM ) + (zA yA + zB yB − 2zM yM )2 − 4(zA zB − z2M )(yA yB − y2M )
2
( √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ )
1
γ 22 = (zA yA + zB yB − 2zM yM ) − (zA yA + zB yB − 2zM yM )2 − 4(zA zB − z2M )(yA yB − y2M )
2
Solving coupling problems using the classical method increases the complexity and computational load with the number of couplings involved. For
example, for the model presented in this paper, a 5-line model (two lines for DC and three lines for AC transmission line) with mutually coupling
between them should be considered, and finally the extracted equations are solved and analyzed for the HPOTL model. The equations presented above
can be applied to any sequence network in turn. It is well documented in the literature that the coupling between positive and negative sequence
networks of lines is very weak. However, coupling between zero sequence networks may be significant and should be the subject of careful analysis
[39]. In [38] positive and zero sequences equivalent circuits are provided for the network presented above in the case of line to ground fault, and the
equivalent circuit can be used to analyze the fault conditions and obtain the voltage and current equations of sending and receiving end of transmission
line in the fault conditions. In this paper, due to the complexity of solving problems related to different faults using analysis of line equations and also
high computational load, software simulations have been used to investigate the effect of different faults on mutually coupled transmission lines.

References [16] J. Liu, J. Duan, H. Lu, Y. Sun, Fault location method based on EEMD and traveling-
wave speed characteristics for HVDC transmission lines, Journal of power and
Energy Engineering 3 (4) (2015) 106–113.
[1] P. Ruffing, C. Petino, S. Rüberg, J.C. Garcia, S. Beckler, A. Arnold, Resonance
[17] Z.-Y. He, K. Liao, X.-P. Li, S. Lin, J.-W. Yang, R.-K. Mai, Natural frequency-based
phenomena and dc fault handling during intersystem faults in hybrid ac/dc
line fault location in HVDC lines, IEEE Trans. Power Delivery 29 (2) (2013)
transmission systems with partial dc cabling. 2018 Power Systems Computation
851–859.
Conference (PSCC), IEEE, 2018, pp. 1–7.
[18] P. Chen, B. Xu, J. Li, A traveling wave based fault locating system for HVDC
[2] E.V. Larsen, R.A. Walling, C.J. Bridenbaugh, Parallel AC/DC transmission lines
transmission lines. 2006 International Conference On Power System Technology,
steady-state induction issues, IEEE Trans. Power Delivery 4 (1) (1989) 667–674.
IEEE, 2006, pp. 1–4.
[3] M. Kizilcay, A. Agdemir, M. Losing, Interaction of a HVDC system with 400-kV AC
[19] P. Chen, B. Xu, J. Li, Y. Ge, Modern travelling wave based fault location techniques
systems on the same tower, in: International Conference on Power System
for HVDC transmission lines, Trans. Tianjin University 14 (2) (2008) 139–143.
Transient in Kyoto, Japan June, 2009, pp. 3–6.
[20] A. Swetha, P.K. Murthy, N. Sujatha, Y. Kiran, A novel technique for the location of
[4] J. J. Hu and B. Bisewski, "Evaluation of coupling between dc and ac transmission
fault on a HVDC transmission line, J. Eng. Appl. Sci 6 (11) (2011) 62–67.
lines on the same right-of-way: parametric analysis and mitigation methods," in
[21] X. Zhang, N. Tai, X. Zheng, W. Huang, Wavelet-based EMTR method for fault
MIPSYCON, 2011.
location of VSC-HVDC transmission lines, The J. Eng. 2019 (16) (2019) 961–966.
[5] "IEEE draft guide for determining fault location on AC transmission and
[22] D. Rezaei, M. Gholipour, F. Parvaresh, A single-ended traveling-wave-based fault
distribution lines," IEEE PC37.114/D9, July 2014, pp. 1-72, 2014.
location for a hybrid transmission line using detected arrival times and TW′ s
[6] M. Abasi, A. Rohani, F. Hatami, M. Joorabian, G.B. Gharehpetian, Fault location
polarity, Electr. Power Syst. Res. 210 (2022), 108058.
determination in three-terminal transmission lines connected to industrial
[23] A. A. Ashagire, "Study on conversion of existing HVAC lines to hybrid HVAC/HVDC
microgrids without requiring fault classification data and independent of line
transmission line to increase transmission capacity and efficiency," Addis Ababa
parameters, Int. J. Electric. Power & Energy Syst. 131 (2021), 107044.
Institute of Technology School of Electrical and Computer Engineering, 2015.
[7] L. Xie, et al., A novel fault location method for hybrid lines based on traveling
[24] A. Novitskiy and D. Westermann, "Some aspects of steady state simulations of AC/
wave, Int. J. Electric. Power & Energy Syst. 141 (2022), 108102, 2022/10/01/.
DC hybrid transmission lines," in 2017 IEEE International Conference on Environment
[8] H. Livani, C.Y. Evrenosoglu, A single-ended fault location method for segmented
and Electrical Engineering and 2017 IEEE Industrial and Commercial Power Systems
HVDC transmission line, Electr. Power Syst. Res. 107 (2014) 190–198.
Europe (EEEIC /I&CPS Europe), 2017, pp. 1-4.
[9] H. Livani, C.Y. Evrenosoglu, A traveling wave based single-ended fault location
[25] C. Petino, P. Ruffing, and A. Schnettler, "Intersystem fault clearing in hybrid AC/
algorithm using DWT for overhead lines combined with underground cables. IEEE
DC power systems with full bridge modular multilevel converters," 2017.
PES General Meeting, IEEE, 2010, pp. 1–6.
[26] G. Ebner, D. Döring, F. Schettler, K. Würflinger, M. Zeller, Fault handling at hybrid
[10] D. Tzelepis, et al., Voltage and current measuring technologies for high voltage
high-voltage AC/DC transmission lines with VSC converters, IEEE Trans. Power
direct current supergrids: a technology review identifying the options for
Delivery 33 (2) (2018) 901–908.
protection, fault location and automation applications, IEEE Access 8 (2020)
[27] X.-h. Li, Z.-p. Liang, Z.-x. Cai, Induced components in HVDC transmission lines
203398–203428.
caused by same-tower AC system faults. 2018 International Conference on Power
[11] A. M. Elhaffar, "Power transmission line fault location based on current traveling
System Technology (POWERCON), IEEE, 2018, pp. 2791–2798.
waves," 2008.
[28] T. Nazarcik and V. Muzik, "Modelling of the mutual influence of the parallel AC/DC
[12] F.H. Magnago, A. Abur, Fault location using wavelets, IEEE Trans. Power Delivery
circuits on the hybrid power transmission line," in 2018 IEEE Conference of Russian
13 (4) (1998) 1475–1480.
Young Researchers in Electrical and Electronic Engineering (EIConRus), 2018, pp. 731-
[13] M. Gilany, E. Eldin, D. Ibrahim, Travelling wave-based fault location scheme for
736.
multi-end aged underground cable system, Int. J. Power and Energy Syst. 28 (2)
[29] J. Prommetta, J. Schindler, J. Jaeger, T. Keil, C. Butterer, AC-protection in the
(2008) 160.
context of AC/Dc-hybrid lines. 2019 IEEE 8th International Conference on
[14] O. Naidu, A.K. Pradhan, A traveling wave-based fault location method using
Advanced Power System Automation and Protection (APAP), IEEE, 2019,
unsynchronized current measurements, IEEE Trans. Power Delivery 34 (2) (2018)
pp. 1854–1858.
505–513.
[30] X. Dong, E. Guan, L. Jing, H. Wang, S. Mirsaeidi, Simulation and analysis of
[15] L. de Andrade, M.T.P. de Leao, Fault location for transmission lines using wavelet,
cascading faults in hybrid AC/DC power grids, Int. J. Electric. Power & Energy
IEEE Latin America Trans. 12 (6) (2014) 1043–1048.
Syst. 115 (2020), 105492.

16
M. Fayazi et al. Electric Power Systems Research 220 (2023) 109302

[31] J. Prommetta, J. Schindler, J. Jaeger, T. Keil, C. Butterer, G. Ebner, Protection [35] M. Ando, E. Schweitzer, R. Baker, Development and field-data evaluation of single-
coordination of AC/DC intersystem faults in hybrid transmission grids, IEEE Trans. end fault locator for two-terminal HVDV transmission lines-part 2: Algorithm and
Power Delivery (2020). evaluation, IEEE Trans. Power Apparatus and Syst. (12) (1985) 3531–3537.
[32] J. Suonan, S. Gao, G. Song, Z. Jiao, X. Kang, A novel fault-location method for [36] A. Greenwood, "Electrical transients in power systems," 1991.
HVDC transmission lines, IEEE Trans. Power Delivery 25 (2) (2009) 1203–1209. [37] M. Szechtman, et al., The CIGRE HVDC benchmark model-a new proposal with
[33] K. De Kerf, et al., Wavelet-based protection strategy for DC faults in multi-terminal revised parameters, Electra 157 (2) (1994) 61–66.
VSC HVDC systems, IET Generation, Trans. Distribution 5 (4) (2011) 496–503. [38] P.M. Anderson, C. Henville, R. Rifaat, B. Johnson, S. Meliopoulos, Power System
[34] C.H. Kim, R. Aggarwal, Wavelet transforms in power systems. Part 1: General Protection, John Wiley & Sons, 2022.
introduction to the wavelet transforms, Power Eng. J. 14 (2) (2000) 81–87. [39] T. Arro and O. Silavwe, "Coupling of transients in HVDC lines to adjacent HVAC
lines and its impact on the AC line protection," 2007.

17