Received 8 February 2023, accepted 1 March 2023, date of publication 6 March 2023, date of current version 14 March 2023.

Digital Object Identifier 10.1109/ACCESS.2023.3253506

Setting Free Fault Location for Three-Terminal

Hybrid Transmission Lines Connected With
Conventional and Renewable Resources
KUKKALA LIKHITHA AND O. D. NAIDU , (Senior Member, IEEE)
Hitachi Energy Grid Automation, Bengaluru, Karnataka 560048, India
Corresponding author: Kukkala Likhitha ([email protected])

ABSTRACT Three-terminal hybrid transmission lines (TTHTLs) are attractive from both environmental and
commercial view. Hybrid transmission lines are growing due to urbanization, connecting an industrial load
and renewable integration. A TTHTL comprises sections of both overhead lines and underground/subsea
cables or overhead lines with different X/R ratios. Faulted section identification (FSI) is key for defining the
adaptive/selective auto-reclosing scheme and estimation of the fault location for TTHTLs. In this paper, FSI
and fault location algorithms are proposed without using the parameters of any line section. Novelty of the
methodology lies in a two-stage approach to the problem. In the first stage, series impedance parameters of all
the sections are calculated using closed loop formulae. These parameters are then utilized in identifying the
faulted section. In the second stage, the above calculated line section parameters and faulted section are used
to estimate complete line parameters including shunt capacitance and subsequently the fault location. The
advantage of the proposed method is that it does not require an initial guess of the line section parameters,
is non-iterative in the first stage, and provides correct fault section identification for TTHTLs. These
features make it suitable for designing the selective auto-reclosing protection scheme for the TTHTLs within
traditional protection relaying hardware. More importantly, the series impedance parameters calculated in
the first stage constitute good initialization values for the non-linear problem of estimating the complete
line parameters. This results in better convergence of the algorithm and accurate parameter estimation.
The developed solution is verified using the PSCAD/EMTDC simulations for TTHTLs connected with
conventional and different inverter-based renewable resources. The performance of the proposed solution
is compared with commercially available solutions, and it is found to be accurate. This solution is amenable
for implementation in line differential protection relays without additional infrastructural changes.

INDEX TERMS Adaptive auto-reclosing protection scheme, inverter-based renewable resources, faulted
section identification, fault location, parameter estimation, three-terminal hybrid transmission lines.

1f 1f 1f
NOMENCLATURE VS , VR ,VT positive sequence during fault voltages
A. VARIABLES measured at terminal S, R, and T
VS
1pre 1pre
, VR
1pre
,VT positive sequence pre-fault voltages respectively.
1f 1f 1f
measured at terminal S, R, and T IS , I R , IT positive sequence during fault currents
respectively. measured at terminals S, R, and T
1pre 1pre 1pre respectively.
IS , I R , IT positive sequence pre-fault currents
1 , Z 1 and Z 1
ZSJ positive sequence impedance of section
measured at terminals S, R, and T RJ TJ
respectively. SJ, RJ and TJ respectively in per km.
R1SJ , R1RJ and R1TJ positive sequence resistance of
The associate editor coordinating the review of this manuscript and section SJ, RJ and TJ respectively in
approving it for publication was Sarasij Das . per km.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.
VOLUME 11, 2023 For more information, see https://creativecommons.org/licenses/by-nc-nd/4.0/ 23839
 K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

1 , L 1 and L 1
LSJ positive sequence inductance of
RJ TJ or overreach [3], [4]. To mitigate these protection issues, the
section SJ, RJ and TJ respectively in line differential is a common protection scheme employed
per km. for three-terminal hybrid lines [5], [6], [7]. However, fault
1 , Y 1 and Y 1
YSJ positive sequence admittance of
RJ TJ section identification, fault location, and parameter estima-
section SJ, RJ and TJ respectively in tion are highly challenging as the line sections have different
per km. surge impedance and line propagation constants. This paper
1 , C 1 and C 1
CSJ positive sequence capacitance of
RJ TJ focus on (i) fault section identification, (ii) fault location, and
section SJ, RJ and TJ respectively in (iii) parameter estimation of the three-terminal hybrid lines.
per km.
1 , Z1
chRJ , ZchTJ
ZchSJ 1 positive sequence characteristic
impedance of section SJ, RJ and TJ B. LITERATURE REVIEW
respectively. Auto-reclosing (AR) is very important for the reliability of
γSJ
1 ,γ1 ,γ1
RJ TJ positive sequence propagation the power system [8]. CIGRE report [9] defines accepted
constant of section SJ, RJ, and TJ practices for ARPS and what is to be practiced in special cases
respectively. like three-terminal hybrid lines. ARPS is enabled for over-
(ABCD)1X ABCD parameters of each section head lines since most faults are transient, but it is blocked for
(X=SJ, RJ, TJ). underground/subsea cables because most of the cable faults
lSJ , lRJ , lTJ length of section SJ, RJ, and TJ are permanent. The guidelines specify that ARPS should be
respectively. used on hybrid lines only if the faulted section is known [9].
For three-terminal hybrid lines, FS identification is challeng-
B. ABBREVIATIONS ing since the parameters such as resistance, inductance is
TTHTLs Three-Terminal Hybrid Transmission Lines. different for each section. Traveling wave [10], [11], [12] and
FSI Faulted Section Identification. AI/ML [13], [14] based faulted section identification methods
AARPS Adaptive Auto-Reclosing Protection Scheme. are proposed for hybrid lines. TW based methods require
ARPS Auto-Reclosing Protection Scheme. high sampling hardware and communication and it is not a
OHL Overhead Line. cost-effective solution. Even though AI/ML technology has
UGC Underground Cable. gained attention in recent literature, the model generality is
SC Subsea Cable. a challenge, and it requires a longer time to mature and for
FL Fault Location. practical deployment of these solutions.
PE Parameter Estimation. Impedance-based fault section identification methods are
TW Traveling Wave. proposed for two-terminal [15] and three-terminal [16] hybrid
IBR Inverter Based Renewable Resources. lines. The accuracy of these methods depends on each sec-
AI Artificial Intelligence. tion’s electrical parameters and these parameters change with
ML Machine Learning. the temperature and aging of the conductors, especially for
TTL Three Terminal Lines. UGC/SC sections. A setting-free fault section identifica-
SCADA Supervisory Control and Data Acquisition. tion is proposed in [17]. This method requires the negative
PMU Phasor Measurement Unit. sequence network equations which may not be available for
IED Intelligent Electronic Device. lines connected with IBRs [18], [19], [20]. Therefore, there
is a requirement for a reliable setting-free FSI algorithm
I. INTRODUCTION for TTHTLs which is used for ARPS and fault location
A. MOTIVATION estimation.
In recent years, several three-terminal hybrid transmis- Locating the fault on the line is important to expedite
sion lines (also known as non-homogeneous or mixed) are the power supply restoration [21]. For quick restoration of
growing in modern power systems due to urbanization, con- faulted transmission lines, an exact FL must be known, else,
necting industrial loads, and renewable integration with min- the maintenance job becomes tiresome and takes a lot of
imum cost. A hybrid transmission line comprises sections of time for power networks spread in rugged geographical ter-
both OHL and UGC/SC or OHL with different X/R ratios, rains [22]. Thus, accurate FL in transmission lines is crucial
to connect offshore wind farms to existing lines, and mitigate for the outage management team to reach the fault point and
right-of-way related issues in urban areas [1]. This reduces start repair at the earliest. Several TW [23], [24], [25] and
the installation and maintenance costs for a substation and impedance-based [26], [27], [28] algorithms are proposed
costs related to the measurement transformers and breaker for two-terminal lines. However, these algorithms cannot
equipment [2]. The distance protection of such configurations be extended directly to three-terminal hybrid lines, as the
is challenging due to the current infeed and outfeed, and impedance of each section is not uniform, and different
different X/R ratios [3], [4] of each section. Current infeed infeed/outfeed of the current lead to more errors in the fault
and outfeed are conditions where the current contribution location estimation [28], [29]. Many fault location algo-
from a line terminal can cause a distance relay to underreach rithms traveling wave based [30], [31], AI/ML [13], [14], and

23840 VOLUME 11, 2023

K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

impedance-based [32], [33], [34], [35], [36], [37], [38], loop as inputs which is difficult for renewable connected
[39], [40] are presented for three-terminal lines including systems. The source impedance of the inverter-based resource
hybrid lines. TW-based methods [30], [31] require each sec- continuously varies during the fault [19] and it may lead
tion’s wave speed as an input, which is not easy to get to larger deviation in fault location calculation. The pro-
in a real-world implementation. Moreover, traveling wave- posed solutions may not work well for networks connected
based methods provide higher accuracy, but it requires high with IBRs. Positive sequence voltage and current-based FL
sampling rates and more communication bandwidth which approaches for TTLs are proposed in [37] and [38]. These
adds to the cost. Although the application of AI/ML in the methods are designed for only homogeneous three-terminal
field of power system monitoring is maturing, for power lines and the accuracy of these methods highly depends on
system protection and fault location applications it may not be the parameters of each section. In [39], proposed a new
practical, as it requires abundant practical or simulated data FSI and FL solution based on synchronized data from all
to train and build the accurate and generic AI/ML models. ends of three-terminal lines. In the absence of data synchro-
Fault location using fundamental phasors are widely used nization, it calculates the value of synchronization operators
because of their ease of use, low sampling rate data require- to form a common reference among the all-end data and
ment, communication bandwidth, and low hardware cost. accuracy depends on the line parameters. In [40], a new
Depending on the measurements required, they are classi- phasor-based technique is proposed to locate faults in non-
fied into single-ended and multi-ended methods [21], [22]. homogeneous/hybrid transmission lines using synchronized
Single-ended methods [26], [27] are easy to implement as it data from all ends. This method uses state estimation to solve
does not require any communication link or data synchroniza- the fault location and they require the impedance matrixes
tion. However, the accuracy of FL greatly varies with sys- of the OHL and UGC sections. Though several methods
tem non-homogeneity, fault resistance, and fault information are reported in the literature, the accuracy of the reported
[28], [29]. These methods will pose larger fault location fault location algorithms depends on the precision of each
errors for lines connected with IBRs [20], as the renewable section parameters. The electrical parameters such as resis-
connected systems introduce larger deviations in phase angles tance, inductance, and capacitance of an electrical line are
of the local and remote currents [19]. Therefore, the perfor- not known with great precision. Many case studies show that
mance of the single-ended approaches is limited for extending actual and stored values of the electrical parameters of the
to the three-terminal hybrid lines due to the infeed/outfeed lines can differ by up to 25–30% [41], especially for UGC
at the tap point and IBR connections. The fault location for section parameters. Line parameter with decent precision
three-terminal double circuit lines is presented using single- is enough for monitoring and protection functions but for
ended data and requires line parameters as a setting [32]. The precise FL, accurate line electrical parameters are essential.
method uses negative sequence quantities to obtain the fault These electrical parameters of each section of TTHTLs are
location. not constant and vary with many weather/seasonal situations
To mitigate the issues of single-ended FL methods, several including the age of the transmission line/cables, etc.
multi-ended fault location methods [33], [34], [35], [36], The estimation of electrical parameters of each section
[37], [38], [39], [40] for three-terminal lines including hybrid will improve the reliability of the FSI and precision of
lines are developed. A negative sequence magnitude-based FL for TTHTLs. Line parameter estimation methods for
method is proposed in [33] which does not require data two-terminal homogeneous transmission lines using a single
synchronization. The FL accuracy does not depend on the set [42], [43] and multiple sets [44], [45] are presented. These
pre-fault load, and it is commercially available. The method methods cannot be applied to the three-terminal hybrid lines,
in [34] computes fault location using the two-ends data of as each section has different parameters and outfeed/infeed
a three-terminal line, the third terminal being connected to at the junction point. A technique [46] is presented for the
renewable energy sources (RES) using positive and nega- estimation of each section electrical parameters for multi-
tive sequence quantities. It requires line parameters as input. terminal lines while the required data are phasor measure-
These lines parameters are varying due to changes in tem- ments at one end of a given line and conventional magnitude
perature, aging, the sag of the conductor, etc., of the line and measurements at the other end. The presented solution is thus
this would affect the accuracy of the fault location. However, based on combining the PMU and SCADA measurements.
the methods [32], [33], [34] may not work for all faults as A non-linear weighted least-square error (NWLSE) algo-
there are no negative sequence currents for balanced faults. rithm is employed for the maximum-likelihood estimation
Moreover, this method will not work for lines connected of parameters. This may not work for hybrid three-terminal
with inverter-based renewable resources as most of the grid lines. To mitigate the need for several data sets, a technique to
codes will not supply negative sequence currents [18], [19]. determine the parameters for TTHTLs using pre and during-
And also, require line parameters as a setting. Fault location fault signals is presented in [47] and [48]. The method in [47]
using only two end measurements [35] and three-terminal uses positive sequence quantities to obtain the fault location
current measurements [36] are proposed for three-terminal along with the line parameters using the trust region optimiza-
lines. These methods require the source impedance and fault tion algorithm. Implementation of such a high computational

VOLUME 11, 2023 23841

 K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

method in the IED platform is difficult. The success of the

method depends on the proper initial guess. The method
in [48] requires positive and negative sequence quantities,
it may not work when IBRs are connected in the network.
These techniques need proper initial guess. The literature
review indicates that there is a scope for research in FSI and
FL for three-terminal hybrid transmission lines without using
any settings.

C. CONTRIBUTIONS
In this paper, a setting-free faulted section identification and
fault location method for three-terminal hybrid transmission
lines connected with conventional/renewable resources using
multi-ended voltage and current signals are proposed. The
method calculates the series impedance of each section by
closed-loop formulae using all terminal pre-fault data. The
faulted section is identified by comparing the calculated and
measured voltage at each terminal using the calculated series
impedance parameters (resistance and inductance) and dur-
ing fault data. This process does not require any complex
iterative computations to identify the faulted section of the
hybrid transmission lines. Therefore, the identified faulted
section information can be used for a selective auto-reclosing
protection scheme (i.e., block for fault in UGC and enable for
fault in OHL) for these hybrid three-terminal lines. The com- FIGURE 1. Shows the geometry of the (a) overhead transmission line and
(b) underground cable, used in the simulation.
plete parameters including shunt capacitance are obtained by
using pre-fault data, using series impedance calculated in the
previous step as an initial guess. As a final step, accurate FL is
multi-ended TW-based methods. Since the proposed method
computed using the estimated electrical constants/parameters
is achieved using the existing hardware of the IED, there is
of each section, faulted section identification information,
no additional cost involved. Further, communication between
and fault data.
the three ends does not require a dedicated communication
The developed solution does not require any user settings
channel. Data is exchanged over IEEE C37.94 protocol and
and reduces the practical deployment time and maintenance
can be shared over a multiplexed communication channel
cost. The main contributions of this paper are as follows:
using Synchronous Digital Hierarchy (SDH) supporting a
• A method for fault location on hybrid three-terminal bandwidth of 2 Mbps. This does not incur any additional cost
transmission line which does not require line parameters for the proposed method, as such communication is essential
to be provided as an input. for working on existing line differential protection IEDs.
• Solution to estimate parameters of the distributed param- Since it does not require additional settings or commissioning
eter (long line corrected equivalent pi) model of the line, to put it in operation, it is practically free of any deploy-
using a single set of voltage and current data, and not ment and overhead costs. Moreover, this algorithm provides
requiring any initial values of the parameters as an input parameters for each section, and those parameters can be used
or user setting. The initial values are instead calculated for other protection and monitoring applications.
through closed loop formulae based on a simplistic RL
model of the line. II. PROPOSED METHOD
• The proposed methodology is deployable in conven- The transmission line is simulated using a frequency-
tional and renewable connected transmission systems. dependent phase model as shown in Fig. 1, which is the most
The developed solution is tested for various fault scenarios realistic representation of a transmission line in simulations.
using PSCAD/EMTDC for 220-kV, 120km main OHL, and However, in our solution approach, in the first stage, we map
20km UGC connected with conventional generations/IBRs. the measurements to a series RL (lumped parameter) model
The simulation results are reliable, and the proposed solution of the line, to avoid complexity. The outcome of this step
is able to identify the faulted section reliably and determine are the estimates of the lumped parameters R and L. In the
the distance to the fault within a 300 - 600 m (two-tower span) next stage, we use these estimates as an initial guess to deter-
distance, using low sampling (1kHz sampling) signals. The mine the parameters namely, characteristic impedance (Zch )
proposed FSI and FL methods are implemented in the existing and propagation constant (γ ) of the distributed parameter
IED/relay platform [7], and it is an economical alternative for (long line corrected equivalent pi) model as shown in Fig. 6.

23842 VOLUME 11, 2023

K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

This approach will eliminate the requirement of multiple data

sets and initial guess of parameters Zch and γ to solve the non-
linear equations that arise in the scenario when we directly try
to map measurements on to the model of Fig. 6. This approach
is simple, requires less computations and easy to implement
in the existing relay platform. The proposed method uses
four modules: (A) setting-free faulted section identification,
(B) calculation of line series impedance, (C) estimation of
line parameters using distributed model, and (D) fault loca-
tion. The details are provided below. FIGURE 3. During fault lumped equivalent model of TTHTLs.

A. FAULTED SECTION IDENTIFICATION METHOD

This section represents the method of identifying faulted
section using the series impedance of lumped model. Let us
consider TTHTLs connected with conventional or inverter-
based resources as shown in Fig. 2.

FIGURE 4. Pre-fault equivalent model of three-terminal lumped model.

positive sequence parameters of section TJ, as in (5),

1f 1f 11f 1 1f
VTcalc = VS − IS ZSJ lSJ + IT ZTJ lTJ (5)
Compute the deviation of calculated voltage from the mea-
FIGURE 2. Three-terminal networks connected to conventional or
sured voltage at each terminal, as in (6) to (8)
renewable power plants. 
 1f

1f 
1VSR = VScalc − VS  (6)
Consider fault on the SJ section at a distance from the fault 
 1f

1f 
of dSJ from Bus S as shown in Fig. 3. Positive sequence volt- 1VRT = VRcalc − VR  (7)
age at Bus S is determined using positive sequence current
 
 1f 1f 
1VTS = VTcalc − VT  (8)
and voltage from Bus R and current from Bus S.
1f
Calculate the junction voltage V JR using Bus R voltage, The minimum the deviation of the calculated voltage from the
current, and impedance parameters, as in (1) measured voltage at each terminal indicates that the sections
1f 1f 1f involved in the computation of the calculated voltage are
V JR = V R − I R Z1RJ l RJ (1)
healthy, concluding that the fault is in another section (in this
Now, using the Bus S current, section SJ impedance param- case-section TJ and RJ are healthy as marked in green in
eters, and estimated junction voltage, compute terminal S Fig. 3).
voltage. The proposed faulted section identification method is sum-
marized as follows,
1f 1f 1f
V Scalc = V JR C I S Z1SJ l SJ (2) • If 1VSR , 1VRT , and 1VTS <∈, then the fault is at the

By substituting equation (1) in (2), we get (3) junction.

• If 1VRT is minimum of 1VSR , 1VRT , and 1VTS , then
1f 1f 1f 1f
V Scalc = V R − I R Z1RJ l RJ C I S Z1SJ l SJ (3) the fault is in the SJ section.
• If 1VTS is minimum of 1VSR , 1VRT , and 1VTS , then
Similarly, voltage at Bus R is determined using current and the fault is in the RJ section.
voltage measured at Bus T, current measured at Bus R, and • If 1VSR is minimum of 1VSR , 1VRT , and 1VTS , then
positive sequence parameters of section RJ, as in (4) the fault is in the TJ section.
1f 1f
1
VRcalc = VT − IT ZTJ
1f 1
lTJ + IR ZRJ lRJ
1f
(4) The accuracy of identifying the faulted section depends on
the series impedance parameters of each section, which are
Finally, voltage at Bus T is determined using current and obtained using pre-fault current and voltage phasors in a
voltage measured at Bus S, current measured at Bus T, and closed-loop formulae and details are provided in section B.

VOLUME 11, 2023 23843

 K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

B. CALCULATION OF SERIES IMPEDANCE PARAMETERS

IN CLOSED LOOP FORMULAE
This section presents a closed loop series impedance param-
eters calculation for TTHTLs with at least one section having
a different set of parameters. Consider R-L model as shown
in Fig. 4. From Fig. 4, KVL across loop I is given as
 
1pre 1pre 1 1pre 1pre 1 1pre
VS − IS ZSJ lSJ − IS + IR ZTJ lTJ − VT = 0
(9)

KVL across loop II is given as

 
1pre 1pre 1 1pre 1pre 1 1pre
VR − IR ZRJ lRJ − IS + IR ZTJ lTJ − VT = 0
FIGURE 6. Pre-fault distributed parameter model (long line corrected
(10) equivalent pi model).

It is assumed that the main section has the same parameters

1 = Z1 .
which are different from the tapped line, i.e, ZSJ from Bus R and T is given in (13) and (14)
RJ
1 is calculated as in (11)
From equations (9) and (10), ZSJ 1f 1f 1f
1
IRJ = −CRJ VR + A1RJ IR (13)
1f 1 1f 1f
1 VS
1pre
− VR
1pre ITJ = −CTJ VT + A1TJ IT (14)
ZSJ = 1pre 1pre
(11)
IS lSJ − IR lRJ From (13) and (14), the current flowing towards the fault
point from the junction is given by (15)
1 in (1), Z 1 is calculated as in (12)
By substituting ZSJ TJ 1f 1f 1f
IJ = IRJ + ITJ (15)
1pre 1pre 1 1pre
1 VS − IS ZSJ lSJ − VT
ZTJ =   (12) Determine the junction voltage from both the healthy termi-
1pre 1pre
IS + IR lTJ nals
1f 1f 1f
The calculated series impedances are used to identify the VRJ = D1RJ VR − B1RJ IR (16)
1f 1f 1f
faulted section described in section -A. The series impedance VTJ = D1TJ VT − B1TJ IT (17)
is also used as an initial guess for the complete parameter
estimation method described in section D. The junction voltage is taken as the average value
1f 1f
1f VRJ + VTJ
C. FAULT LOCATION METHOD FORMULATION VJ = (18)
2
This section presents the method of obtaining distance to
where
fault for hybrid three-terminal lines. The formulation of fault s s
location [37] is obtained by equating the voltage at the fault 1
ZRJ R1RJ + jωLRJ
1
1
point determined from both terminals of the faulted section. ZchRJ = 1
= 1
;
YRJ jωCRJ
The detailed steps are as follows. q q
γRJ
1
= 1 Y 1 = (R1 + jωL 1 )(jωC 1 );
ZRJ RJ RJ RJ RJ
   
A1RJ = D1RJ = cosh γRJ 1
lRJ ; B1RJ = ZchRJ
1
sinh γRJ
1
lRJ ;
sinh γRJ1 l


1 RJ
CRJ = 1
;
ZchRJ
(ABCD)1RJ are ABCD parameters of section RJ
s s
1
ZTJ R1TJ + jωLTJ1
1
ZchTJ = 1
= 1
;
YTJ jωCTJ
FIGURE 5. A-g fault on SJ section for a TTHTLs. q q
γTJ
1
= ZTJ 1 Y 1 = (R1 + jωL 1 )(jωC 1 );
TJ TJ TJ TJ
Consider fault on SJ section at a distance of d SJ from Bus S    
as shown in Fig. 5 and corresponding pre and during fault net- ATJ = DTJ = cosh γTJ lTJ ; BTJ = ZchTJ sinh γTJ
1 1 1 1 1 1
lTJ ;
work equivalent are shown in Fig. 6 and Fig. 7 respectively.
sinh γTJ 1 l


1 TJ
Calculate the current flowing towards the junction from both CTJ = ;
1
ZchTJ
the healthy sections. The current flowing towards the junction

23844 VOLUME 11, 2023

K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

FIGURE 7. During fault distributed parameter model for fault in section SJ at a distance of d SJ from Bus S.

(ABCD)1TJ are ABCD parameters of section TJ location as in (21)

Determine voltage at the fault point using current and 1

K1

voltage measured from Bus S dSJ = tanh−1 (21)
γSJ
1 K2
1f 1f 1f
VSF = D1dSJ VS − B1dSJ IS (19) The inverse hyperbolic tangent cannot be implemented in
IED platforms directly; therefore equation (21) can be further
Determine voltage at the fault point using current and voltage simplified as shown in equation (22) which will be easy to
measurements obtained at the junction from previous steps implement.
    
1f 1f
VJF = D1lSJ −dSJ VJ − B1lSJ −dSJ IJ
1f
(20) 0.5 K1 K1
dSJ = 1 ln 1 + − ln 1 − (22)
γSJ K2 K2
where where,
s s
1 1f 1f 1f
1 ZSJ R1SJ + jωLSJ
1
K 1 = VS − D1SJ D1RJ VR + D1SJ B1RJ IR
ZchSJ = = ;
Y1 jωCSJ1
− B1SJ CRJ
1 1f
VR + B1SJ A1RJ IR
1f
q SJ q
1f 1f
γSJ
1
= ZSJ 1 Y 1 = (R1 + jωL 1 )(jωC 1 )
SJ SJ SJ SJ − B1SJ CTJ
1
VT + B1SJ A1TJ IT (23)
1 1f 1 1 1f 1 1 1f 1f
BRJ IR − D1SJ CRJ
1
   
A1dSJ = DdSJ = cosh γSJ dSJ ;BdSJ = ZchSJ sinh γSJ
1 1 1 1 1
dSJ ; K2 = ZchSJ (IS − CSJ DRJ VR + CSJ VR
1f 1f 1f
sinh γSJ 1 d

+ D1SJ A1RJ IR − D1SJ CTJ
1
VT + D1SJ A1TJ IT ) (24)
SJ
Cd1SJ = 1
;
ZchSJ Similarly, for fault in the RJ section at a distance of dRJ from
Bus R, the unknown fault location dRJ is obtained as in (25),
(ABCD)1dSJ are ABCD parameters of the section, from Bus S 0.5
 
K1
 
K1

to the fault point dRJ = 1 ln 1 + − ln 1 − (25)
γRJ K2 K2
 
A1lSJ −dSJ = D1lSJ −dSJ = cosh γSJ 1
(lSJ − dSJ ) ; where,
1f 1f 1f
K 1 = VR − D1RJ D1SJ VS + D1RJ B1SJ IS
 
B1lSJ −dSJ = ZchSJ sinh γSJ 1
(lSJ − dSJ ) ;
1f 1f
− B1RJ CSJ
1
VS + B1RJ A1SJ IS
sinh γSJ 1 (l − d )


1 SJ SJ
ClSJ −dSJ = ; 1f 1f
1
ZchSJ − B1RJ CTJ
1
V + B1RJ A1TJ IT (26)
 T
1 1f 1 1f 1 1 1f
K 2 = ZchRJ ∗ IR − CRJ D1SJ VS + CRJ BSJ IS
(ABCD)1lSJ −dSJ are ABCD parameters of the section, from
1f 1f
fault point to the junction − D1RJ CSJ
1
VS + D1RJ A1SJ IS
By equating the voltage at the fault point when calcu-

1f 1f
− D1RJ CTJ
1
VT + D1RJ A1TJ IT (27)
lated from Bus S and Junction, we obtain the unknown fault

VOLUME 11, 2023 23845

 K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

Similarly, for fault in the TJ section at a distance of dTJ from The non-linear objective functions are solved using the least-
Bus T, the unknown fault location dTJ is obtained as in (28), squares iterative method, with an initial guess as calculated
0.5
 
K1
 
K1
 in section -B,
dTJ = 1 ln 1 + − ln 1 − (28)
γTJ K2 K2  T
X0 = R1SJ , LSJ
1
, 1.5CSJ
1
, R1TJ , LTJ
1
, 1.5C 1TJ
where,
1f 1f 1f This results in better convergence of the algorithm and accu-
K 1 = VT − D1TJ D1SJ VS + D1TJ B1SJ IS
1f 1f rate parameter estimation. These estimated parameters are
− B1TJ CSJ
1
VS + B1TJ A1SJ IS used to calculate precise fault location as formulated in
1f 1f
− B1TJ CRJ
1
V + B1TJ A1RJ IR (29) section C. This approach is setting-free and it does not require
 R additional hardware infrastructure or additional engineering
1 1f 1 1f 1 1 1f
K 2 = ZchTJ ∗ IT − CTJ D1SJ VS + CTJ BSJ IS costs. Fig. 8 depicts the summary of the proposed solution.
1f 1f
− D1TJ CSJ
1
VS + D1TJ A1SJ IS

1f 1f
− D1TJ CRJ
1
VR + D1TJ A1RJ IR (30)
The fault location obtained for each section requires line
parameters including capacitance of the three-terminal hybrid
transmission line. The required line parameters are estimated
using the pre-fault data in section D.

D. ESTIMATION OF COMPLETE LINE PARAMETERS

In this section, the parameter estimation formulae are derived
using pre-fault network equivalent as shown in Fig. 6. The
objective functions are formulated as follows:
By equating junction voltage when calculated from Bus S
and T measurements, we get FI as
1pre 1pre 1pre 1pre
FI = D1SJ VS − B1SJ IS − D1TJ VT + B1TJ IT =0
(31)
Similarly, by equating junction voltage when calculated from
Bus R and T measurements, we get FII as
1pre 1pre 1pre 1pre
FII = D1RJ VR − B1RJ IR − D1TJ VT + B1TJ IT =0
(32)
Finally, by applying KCL at the junction, we get FIII as
1 1pre 1pre 1pre 1pre
FIII = CSJ VS − A1SJ IS 1
+ CRJ VR − A1RJ IR FIGURE 8. Flowchart of the proposed solution.
1 1pre 1pre
+ CTJ VT − A1TJ IT =0 (33)
where III. RESULTS AND ANALYSIS
    The results for lines connected with conventional/inverter-
A1SJ = D1SJ = cosh γSJ
1
lSJ ; B1SJ = 1
ZchSJ sinh γSJ
1
lSJ ;
based renewable resources are presented in this section, along
sinh γSJ
1 l )


1 SJ with illustrative examples. A detailed investigation of the
CSJ = 1
; developed solution with numerous power system fault situa-
ZchSJ
tions is also presented. The impact of change in each section’s
(ABCD)1SJ are ABCD parameters of section SJ. electrical and source parameters on practically existing FL
There are a total of six unknowns to be estimated. methods along with the proposed solution is also discussed
 T
X = R1SJ , LSJ 1
, CSJ
1
, R1TJ , LTJ
1
, CTJ
1 in this section.

The objective functions are further split into real and imagi- A. ILLUSTRATIVE EXAMPLE FOR TTHTL CONNECTED WITH
nary parts. CONVENTIONAL SOURCES AT ALL TERMINALS
f1 (X ) = Re(FI ) 

A detailed analysis of the developed method has been pre-
f2 (X ) = Im(FI ) 


 sented in this section. COMTRADE99 format is used to
f3 (X ) = Re(FII )

 record the fault data. The algorithm has been verified through
(34)
f4 (X ) = Im(FII )  MATLAB scripts. The sampling rate used in this method is
f5 (X ) = Re(FIII ) 


 1kHz, the phasors are obtained through the DFT technique.
f6 (X ) = Im(FIII )

The proposed method has been tested for 220kV, 50 Hz


23846 VOLUME 11, 2023

K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

TABLE 1. Determined electrical parameters of each section of TTHTL. TABLE 2. Estimated line parameters using distributed line model.

TABLE 3. Identified faulted section and fault location.

system, with the length of each section being: SJ-80km,
RJ-40km, and TJ-20km.
Case 1- AB fault on SJ section at 40 km from Terminal
S with RF 10: The test system considered for this case is
derived by closing switch S1 in Fig. 2, all the terminals are
connected to conventional sources. Consider the AB fault
on the SJ section at 40km from terminal S on the OHL
with a fault resistance of 10. The current and voltage sig- Section II-B and the line parameters estimated from
nals measured at Bus S, Bus R, and Bus T are shown in Section II-D are well matched to the actual parameters. The
Fig. 9(a), (b), and (c) respectively. identified faulted section and calculated fault location are
presented in Table 3. The faulted section is identified cor-
rectly using the calculated series impedance parameters. The
estimated parameters are utilized to obtain the FL as they are
highly precise than the determined parameters using closed
loop formulae. The fault location obtained using the during
fault phasors and estimated line parameters is 40.02 km. The
obtained FL error using (35) is 0.02%.
%Error
 
 Actual Fault Location-Calculated Fault Location 
=  ×100
Total line length 
(35)

B. STUDY FOR TTHTL CONNECTED WITH

INVERTER-BASED RENEWABLE RESOURCES
This section presents illustrative example cases for the valid-
ity of the proposed method with different inverter-based
resources such as wind type IV and solar PV plants. The wind
and solar modeling details are provided below for this study.

1) MODELING OF INVERTER-BASED RENEWABLE

RESOURCES (WIND TYPE IV)
The 200 MW generator consists of 100 units and each unit
is of 2 MW rated power. The output of the generator is at
0.69 kV which is stepped up to the collector level voltage
FIGURE 9. Voltages and currents recorded at (a) Bus S, (b) Bus R, and
(c) Bus T, for AB fault at 40km from Bus S.
of 33kV and then further to the PCC voltage of 220 kV.
The Type IV wind turbine generator is a full converter-based
The calculated series impedance in a closed loop form and generator where the whole power is evacuated to the grid
estimated line parameters are obtained using pre-fault data. through back-to-back converters (Fig. 10(a)) [49]. Various
For this case, the respective error of each parameter when control strategies to maximize the power delivery of the grid
compared to actual electrical parameters of each section are connected IBR have been proposed [50], [51]. The schematic
provided in Tables 1 and 2 respectively. From the tables, of the control strategy which is achieved in a synchronous
it is observed that the series impedance calculated from reference frame is provided in Fig. 10(b) [52]. The outer

VOLUME 11, 2023 23847

 K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

FIGURE 10. Wind Type IV generator modeling details (a) detailed control scheme, and (b) schematic of the grid side inverter control.

TABLE 4. Source Impedances for conventional sources. TABLE 5. Simulation parameters of inverters.

control loop generates q axis current reference aimed at main-

taining PCC voltage magnitude at a reference of 1pu (voltage
control). Further, d axis current reference is generated aimed
at maintaining the DC link voltage at the reference of 1pu.
The current limiter is applied on the q-axis and d-axis current
references generated from the outer loop, to limit the total
current to the set limit of 1.1pu. Further, the inner control
loop generates the converter reference voltages for the grid-
side converter. Feedforward compensation to decouple the
wind generator from the grid is also included. This ensures
balanced voltage across the filter, even during unbalanced
conditions at the grid end. In addition to these, the fault ride
through (FRT) capability is also implemented which ensures
additional reactive current injection during low voltage con-
ditions. A gain (k) of 2 (which determines the rate of reactive
FIGURE 11. Fault ride through (FRT) characteristic.
current injection in proportion to the dip in PCC voltage) and
a dead band of 0.1 is provided (which means the additional
injection starts only if the dip in voltage is greater than 0.1pu) to conventional sources and terminal T is connected to a type
(Fig. 11) [49]. However, these FRT requirements depend on IV Wind farm. The current and voltage signals measured at
the grid code [18]. The parameters of the wind generator are Bus S, Bus R, and Bus T are shown in Fig. 12(a), (b), and (c)
summarized in Table 5. respectively. From Fig. 12, the current magnitude is limited
and modulated (appears like a balanced fault (see Fig. 12(c))
2) ILLUSTRATIVE CASES—THREE TERMINAL HYBRID LINES for IBR connected terminal due to different converter control
CONNECTED WITH RENEWABLE RESOURCES and grid codes followed by the IBR. Despite such unusual
In this section, we simulated different cases for three-terminal behavior of the IBR-connected terminal current signals, the
hybrid lines connected with IBRs. The analysis is provided proposed method provides accurate results. For this case,
below. the calculated series impedance in a closed loop form, esti-
Case 1- A-g fault on TJ section at 10 km from Bus T with mated electrical parameters of each section of TTHTL, and
RF 2 : The test system considered for this case is derived by their errors with respect to the actual electrical parameters
closing switch S2 in Fig. 2, terminals S, and R are connected of each section are provided in Tables 6 and 7 respectively.

23848 VOLUME 11, 2023

K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

TABLE 6. Determined electrical parameters of each section of TTHTL.

TABLE 7. Estimated line parameters using distributed line model.

FIGURE 12. Voltages and currents measured at (a) Bus S, (b) Bus R and
(c) Bus T, for A-g fault at 10km from Bus T.

The identified faulted section and calculated fault location are

presented in Table 8. The calculated series impedance iden-
tifies the faulted section accurately. Precise FL is obtained
using the estimated electrical parameters of each section of
TTHTL for this case. Integration of the wind farm at Bus T TABLE 8. Identified faulted section and fault location.
did not affect the proposed method and provided accurate
results. The error in fault location obtained for this case is
0.02% which is within two tower span distance.
Case 2- BC-g fault on RJ section at 110 km from Bus
S with RF 20 : The test system considered for this case
is derived by closing switch S3 in Fig. 2, terminals S, and
TABLE 9. Determined electrical parameters of each section of TTHTL.
R are connected to conventional sources and terminal T is
connected to the solar power plant. The solar park is mod-
eled as a full converter model [53]. The current and voltage
signals measured at Bus S, Bus R, and Bus T are shown in
Fig. 13(a), (b), and (c) respectively. For this case, the deter-
mined series impedance in a closed loop form, estimated elec-
trical parameters of each section of TTHTL, and their errors
with respect to the actual section parameters are presented
in Tables 9 and 10 respectively. The identified faulted section
and calculated fault location are presented in Table 11. The
table shows the calculated/estimated electrical parameters of
each section accurately even for the solar park connected
at terminal T. The faulted section and location are achieved
correctly for these cases using the estimated parameters. The measured at Bus S, Bus R, and Bus T are shown in Fig. 14(a),
error in fault location obtained for this case is 0.48%. The (b), and (c) respectively. From Fig. 14, reduced the currents
integration of renewable sources like solar PV, has the least at terminal R and another terminal are increased. The current
impact on the proposed method as it uses all side measure- signal is modulated at terminal T due to inverter controls.
ments [20]. The non-homogeneity introduced by IBRs are The junction faults are very common for the three-terminal
compensated [43] by using all side measurements. hybrid lines. For this case, the calculated series impedance
Case 3- A-g fault at Junction with RF 50 : The test in a closed loop form, estimated electrical parameters of
system considered for this case is derived by closing switch each section, and their deviations with respect to the actual
S2 and S3 in Fig. 2, terminals S, and R are connected to parameters of each section are presented in Tables 12 and 13.
conventional sources and terminal T is connected to a type From the table, OHL section parameters errors are within 1%.
IV Wind farm and Solar PV. The current and voltage signals The identified faulted section and calculated fault location

VOLUME 11, 2023 23849

 K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

FIGURE 14. Voltage and current signals measured at (a) Bus S, (b) Bus R
FIGURE 13. Voltages and currents measured at (a) Bus S, (b) Bus R and and (c) Bus T for A-g fault at junction.
(c) Bus T, for BC-g fault at 110km from Bus S.

TABLE 12. Determined electrical parameters of each section of TTHTL.

TABLE 10. Estimated line parameters using distributed line model.

TABLE 13. Estimated line parameters using distributed line model.

TABLE 11. Identified faulted section and fault location.

are presented in Table 14. The faulted section is identified

as a junction fault for this case as all calculated and measured
voltage differences are less than the threshold.
The fault location provided for this case is section lengths TABLE 14. Identified faulted section and fault location.
from each terminal. For example, IED at terminal S provides
the fault location is 80km.
Case 4-BC-g fault on RJ section at 105 km from Bus S with
RF 20  : The test system considered for this case is shown
in Fig. 15, terminal S is connected to a conventional source,
R is connected to the solar power plant and T is connected
to a type IV Wind farm. This arrangement might not be
feasible today, however, this could be a possible scenario for signals measured at Bus S, Bus R, and Bus T are shown in
future renewable integration systems. The current and voltage Fig. 16(a), (b), and (c) respectively.

23850 VOLUME 11, 2023

K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

TABLE 15. Determined electrical parameters of each section of TTHTL.

FIGURE 15. Three-terminal hybrid line connected to conventional, wind

farm and solar plants.
TABLE 16. Estimated line parameters using distributed line model.

TABLE 17. Identified faulted section and fault location.

TABLE 18. Various fault conditions and test system configurations.

FIGURE 16. Pre and fault voltages and currents measured at (a) Bus S,
(b) Bus R and (c) Bus T for BC-g fault at 105km from Bus S.

For this case, the calculated series impedance in a closed

loop form, estimated line parameters, and their corresponding
errors with respect to the actual line parameters are provided
in Tables 15 and 16 respectively. The identified faulted sec-
tion and calculated fault location are presented in Table 17.
From table 16, the estimated parameter is accurate and the
error in inductance estimation is less than 1%. The faulted
section is accurately identified as ‘‘RJ’’ and the location
is calculated as 105.08km. The fault location error is less
than 0.1%. The proposed solution is not affected even by
two terminals connected with IBRs. The proposed solution
is suitable for future power system scenarios.

C. COMPREHENSIVE STUDY USING DIFFERENT TEST

SYSTEMS AND CONDITIONS
The proposed solution is tested for various combinations of are tested; the average and maximum errors of parameters
the IBRs with various power system fault conditions. The test of each section are provided in Table 19. From the table,
cases are presented in Table 18. A total of 4320 fault cases the average error is less than 2% which is suitable for fault

VOLUME 11, 2023 23851

 K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

TABLE 19. Average and maximum error of estimated line parameters. TABLE 20. Average and maximum error of estimated line parameters of
test system 1.

TABLE 21. Average and maximum error of estimated line parameters of

test system 2.

FIGURE 17. Performance summary of the developed FL solution for

various fault situations.
TABLE 22. Reliability of FSI and FL error (%) for test system 1 and 2.

FIGURE 18. Practical three-terminal systems (a) Test System 1 and

(b) Test System 2.

section SJ. This fault is identified as a Junction fault as section

location and protection setting calculations. The faulted sec- SJ is very short (0.12km). The average fault location error for
tion is identified correctly for all the cases and the fault both the test systems is close to 300m (i.e., close to two tower
location error is demonstrated in Fig. 17. From Fig. 17, the span distance) which is the best accuracy as per the industry
maximum fault location error is 0.68%. The average fault standards for these complex systems.
location error is 0.2% (∼240m) for about 3450 (∼80%) cases
which is within two tower span (300m) distance. The fault E. COMPARATIVE ANALYSIS WITH COMMERCIALLY
location is not influenced by the fault type, fault resistance, AVAILABLE METHODS FOR TTHTLs
loading of the line, SIRs (line to source impedance ratio), and This section compares the performance of the proposed solu-
location of the inverter-based renewable resource. tion with practically proven positive [38] and negative [33]
sequence-based methods. These methods require the line
D. PERFORMANCE EVALUATION FOR PRACTICAL parameters as input. The accuracy of fault location highly
INDIAN SYSTEMS depends on the parameters of each section such as resistance,
The proposed method is tested with two practical systems as inductance, and capacitance. These parameters are never
shown in Fig. 18. These lines are connected to a 110kV, 50Hz accurate most of the time, as these parameters change with
supply. A total of 144 cases are tested for each test system temperature, aging, the sag of the conductor, etc., of the line,
covering different fault types, fault resistance, fault inception especially in cable sections. A comparison of the proposed
angles, and locations of the fault. The parameter estimation method with commercially established positive sequence
algorithms are accurately working for these test systems even (PSFL) [38] and negative sequence (NSFL) [33] methods for
for a short line with taps. The average and maximum errors in lines connected with conventional and IBRs is presented in
parameter estimation are provided in Tables 20 and 21. From this section.
the tables, the proposed method calculated the parameters
accurately for Test Systems 1 and 2. The faulted section 1) THREE TERMINAL HYBRID LINES CONNECTED WITH
identified for Test Systems 1 and 2 and fault locations for CONVENTIONAL SOURCES
the cases are provided in Table 22. The faulted section is This section provides the comparative assessment of the line
identified accurately for both test systems. Two cases failed and source impedance parameter variation on the proposed
for Test System 2 as the fault is very close to the junction on method (PM), positive sequence (PSFL) [38], and negative

23852 VOLUME 11, 2023

K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

sequence (NSFL) [33] for lines connected with conventional the variation in source impedance magnitude. With the inte-
sources. gration of renewables into the grid, the source strength varies
enormously, and the accuracy of the method [33] will only
a: VARIATION IN EACH SECTION PARAMETERS get worse for networks with integrated renewable sources.
Consider a phase-to-ground (A-g) fault at 50km from termi-
nal SJ. The fault resistances considered for the fault are 20 
and the FIA is 00 . By varying the line parameters (resistance,
inductance, capacitance) with error of 0% to 50%, the fault
location errors for the proposed method (PM) and practical
methods are evaluated. The evaluation result for the fault case
is shown in Fig. 19. From the figure, it is observed that the
fault location errors of PSFL [38] and NSFL [33] increases
with the increase in the error in section parameters whereas
the FL error for PM is not impacted by the parameter variation
in this case. For a 10% variation in the line parameters, the FL
error for practical methods is more than 3%. This requires
a significant amount of time, and more towers need to be FIGURE 20. Comparative assessment of the developed vs commercially
established solutions for change in the source equivalent magnitude for
inspected for the existing methods. The accuracy of the main A-g fault at 15km from Terminal T.
and tapped lines parameters has a significant impact on the
FL accuracy claimed by existing methods.
2) THREE TERMINAL HYBRID LINES CONNECTED WITH
INVERTER-BASED RENEWABLE SOURCES
This section provides the comparative assessment of the
line and source impedance parameter variation on the pro-
posed method (PM), positive sequence (PSFL) [38], and
negative sequence (NSFL) [33] for lines connected with
inverter-based renewable resources.

a: INVERTER-BASED RESOURCES CONNECTED AT

TAPPED LINE TERMINAL T
FIGURE 19. Comparative analysis of the developed solution vs The test system considered for this case is derived by closing
commercially established solutions for change in the parameters of
TTHTLs for A-g fault at 50km from Terminal S. switch S2 in Fig. 2, terminals S, and R are connected to con-
ventional sources and terminal T is connected to wind type IV.
With the integration of renewable power plant in the present
b: VARIATION IN SOURCE PARAMETERS scenario, this configuration is very common. Let us examine
The presented solution using negative sequence quanti- how effectively the existing and proposed method behaves for
ties [33] requires equivalents of sources magnitudes param- this configuration. Consider an A-g fault on the transmission
eters as a setting. The magnitudes of the source equivalent line at 75 km from terminal S having a fault resistance of
change with changes in operational and structural conditions 10 , with a fault inception angle of 0◦ .The faulted section is
in the interconnected power system. A comparative assess- correctly identified by the proposed and positive sequence-
ment of the developed and practically available FL solution based method [38]. However, the positive sequence-based
for source impedance magnitude changes, is presented in this method [38] requires each section parameter as an input. The
section. For this case, we have considered a phase-to-ground calculated fault distances using PSFL [38] and the proposed
(A-g) fault on tapped line (i.e., 15km from Terminal T). The method are 75.8 km and 74.72 km respectively. The absolute
fault resistance for A-g fault is 1 as practical fault resistance percentage of fault location errors of PSFL and PM are 0.6%
on cable is very small due to the small arc length in between and 0.2% respectively for this case. The proposed and positive
the core and sheath. sequence methods are not affected by renewable integration.
By varying the error in magnitude of source impedance However, the performance of PSFL [38] highly depends on
from 0% to 50%, the performance of the proposed method the precision of the parameters of the line.
is validated. The results of the analysis for these cases are The solution presented using negative sequence quanti-
shown in Fig. 20. From the figure, PSFL [38], and proposed ties [33] is not worked for this method as the negative
method fault location accuracies are not impacted by the sequence current is not available at Terminal T as shown in
source parameter variation for this case, but the fault loca- Fig. 21. NSFL [33] will not provide reliable results for lines
tion error of negative sequence [33] increases with increase connected with inverter-based renewable resources (Solar PV
in error of source impedance magnitudes parameters. The or Wind type IV) as most of the grid codes will not provide the
negative sequence-based method [33] is greatly affected by negative sequence currents [18], [19]. Moreover, the source

VOLUME 11, 2023 23853

 K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

TABLE 23. A comparative analysis with existing approaches.

impedance of the inverter-based resource is continuously

varying during the fault period and the method [33] is not
reliable for renewable connected systems.

FIGURE 22. Shows negative sequence current measured at all terminals

for AB fault at 5km from terminal T.

sequence-based method [33] does not work as the negative

FIGURE 21. Shows recorded negative sequence quantities at all terminals sequence currents are not available at terminals T and R as
for A-g fault at 75km from terminal S.
shown in Fig. 22.

b: INVERTER-BASED RESOURCES CONNECTED AT MAIN F. QUALITATIVE ANALYSIS OF RECENT LITERATURE

LINE AT TERMINAL R AND TAPPED LINE TERMINAL T The proposed method is compared with the practically proven
The test system considered for this case is shown in Fig. 15. methods [33], [38] and found to be superior with various
The Type IV wind farm is connected at tapped line Terminal T fault types, fault resistance, source impedances, and differ-
and solar PV is connected at Terminal R. This type of con- ent network configurations. Further, the proposed method is
figuration may not be common today, but it may be possi- compared with recent methods [34], [47]. The method in [34]
ble in future renewable connected systems. Let us examine computes fault location using the two-ends data of a three-
how effectively the existing and proposed method behaves terminal line, the third terminal being connected to renewable
for this configuration. Consider AB fault on the hybrid line energy sources (RES) using positive and negative sequence
at 5km from Terminal T having a fault resistance of 2 quantities. Method [34] requires negative sequence quantities
with a fault inception angle of 0◦ . The faulted section is which are not available for balanced faults, and it requires line
correctly identified by proposed and positive sequence [38] parameters as input. These line parameters are varying due to
methods. The estimated fault distances using PSFL [38] and changes in temperature, aging, the sag of the conductor, etc.,
the proposed method are 5.4km and 4.68km respectively. of the line and this would affect the accuracy of the fault loca-
The absolute percentage of fault location errors of PSFL and tion. The method in [47] uses positive sequence quantities to
PM are 0.33% and 0.26% respectively for this case. The obtain the fault location along with the line parameters using
proposed and positive sequence methods are not affected by the trust region optimization algorithm. Implementation of
renewables connected with two ends of three- NSFL [33] such a high computational method in the IED platform is
failed to calculate the faulted section and fault location for difficult. The success of the method depends on the proper
the lines connected with renewable resources. The negative initial guess. The proposed method overcomes limitations

23854 VOLUME 11, 2023

K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

of [34] and [47], as it computes initial guess in a closed- In future work, the developed solution can be implemented in
loop form which makes it computationally less intensive. It is cloud-based platforms or SCADA/EMS control centers. The
also advantageous in comparison to the method in [34], as it issues with measurement errors and data synchronization can
does not require line parameters as settings and works for all be eliminated in future works.
fault types. A detailed comparison is provided in Table 23 by
taking into consideration the factors that affect the accuracy REFERENCES
of fault location. This will help to choose the appropriate fault [1] Protection Aspects of Multi-Terminal Lines, IEEE Power System Relaying
Committee (PSRC), New York, NY, USA, 1980.
locator for practical usage by utility providers.
[2] Protection Aspects of Multi Terminal Lines, document TH0056-2-PWR,
IEEE Power System Relaying Committee, 1979.
[3] D. A. Tziouvaras and J. Needs, ‘‘Protection of mixed overhead and under-
IV. CONCLUSION ground cable lines,’’ in Proc. 12th IET Int. Conf. Develop. Power Syst.
In this work, a setting-free fault section identification and Protection (DPSP), 2014, pp. 1–6.
fault location estimation solution using all terminal voltages [4] S. Sarangi and A. K. Pradhan, ‘‘Adaptive direct underreaching transfer trip
protection scheme for the three-terminal line,’’ IEEE Trans. Power Del.,
and currents for a TTHTL connected with conventional or
vol. 30, no. 6, pp. 2383–2391, Dec. 2015.
IBRs is presented. The main advantages of the proposed [5] R. K. Aggarwal and A. T. Johns, ‘‘The development of a new high speed
method are summarized as follows. 3-terminal line protection scheme,’’ IEEE Power Eng. Rev., vol. PER-6,
no. 1, p. 43, Jan. 1986.
• The proposed method eliminates the requirement of line [6] N. Jin, X. Lin, J. Xing, Z. Li, S. Liu, L. Chen, X. Ma, Z. Rong, and N. Tong,
‘‘Research on multiterminal current differential protection criterion with
parameters of each section as settings, which enhances high sensitivity and synchronization error tolerance capability,’’ IEEE
the accuracy of the solution and reduces engineering cost Trans. Power Del., vol. 33, no. 6, pp. 3085–3094, Dec. 2018.
and time. [7] Hitachi Energy Product Guide. Line Differential Protection RED6702.1.
Accessed: Mar. 2019. [Online]. Available: https://search.abb.com/library/
• One of the main contributions of the proposed method Download.aspx?DocumentID=1MRK505346-BEN&LanguageCode=en
is that the fault section is identified using closed-loop &DocumentPartId=&Action=Launch
formulae which can be used for selective/adaptive auto- [8] IEEE Guide for Automatic Reclosing of Circuit Breakers for AC Distri-
bution and Transmission Lines, IEEE Standard C37.104-2012, Jul. 2012,
reclosing protection schemes for hybrid lines. doi: 10.1109/IEEESTD.2012.6232415.
• The performance of the proposed solution correctly [9] Short Circuit Protection of Circuits With Mixed Conductor Technologies
identifies the faulted section and locates the fault in Transmission Networks, CIGRE, Paris, France, Jun. 2014.
[10] N. George and O. D. Naidu, ‘‘Traveling wave based autoreclosure scheme
accurately for TTHTLs connected with conventional/ for multi-terminal lines,’’ in Proc. IEEE PES Innov. Smart Grid Technol.
inverter-based renewable resources for various fault sit- Conf. Eur., Sep. 2017, pp. 1–6.
uations including various grid codes. [11] O. D. Naidu, A. K. Pradhan, and P. Krishnamurthy, ‘‘Traveling wave
based adaptive auto-reclosing and fault location for three-terminal mixed
• The proposed method is compared with the commer- lines,’’ in Proc. IEEE PES GTD Grand Int. Conf. Expo. Asia, Mar. 2019,
cially existing methods and found to be superior with pp. 466–471.
various fault types, fault resistance, source impedances, [12] R. J. Hamidi and H. Livani, ‘‘Traveling-wave-based fault-location algo-
rithm for hybrid multiterminal circuits,’’ IEEE Trans. Power Del., vol. 32,
and different network configurations. no. 1, pp. 135–144, Feb. 2017.
• The proposed method is tested with two practical Indian [13] H. Livani and C. Y. Evrenosoglu, ‘‘A machine learning and wavelet-based
systems and the method works well for these systems. fault location method for hybrid transmission lines,’’ IEEE Trans. Smart
Grid, vol. 5, no. 1, pp. 51–59, Jan. 2014.
• Obtained results show that, with various fault scenar- [14] A. Tabatabaei, M. R. Mosavi, and P. Farajiparvar, ‘‘A traveling-wave fault
ios, the fault section identification achieved 100% reli- location technique for three-terminal lines based on wavelet analysis and
ability, and the average FL precision is around 0.2%. recurrent neural network using GPS timing,’’ in Proc. Smart Grid Conf.
(SGC), Dec. 2013, pp. 268–272.
It is predictable that the accuracy of the FL will be [15] S. Zhang, H. Gao, and Y. Song, ‘‘A new fault-location algorithm for extra-
within 2 tower spans for TTHTLs connected with high-voltage mixed lines based on phase characteristics of the hyperbolic
conventional/IBRs. tangent function,’’ IEEE Trans. Power Del., vol. 31, no. 3, pp. 1203–1212,
Jun. 2016.
• Another interesting attribute of the proposed method is [16] O. Naidu, P. Yalla, and N. George, ‘‘Auto-reclosing protection scheme for
that it does not need additional infrastructure other than multi-terminal mixed lines using synchrophasor measurements,’’ in Proc.
what is needed for a line differential protection scheme, Int. Conf. Smart Grid Synchronized Meas. Anal. (SGSMA), May 2019,
pp. 1–6.
i.e., a 2 Mbps communication link, and no additional [17] O. D. Naidu, N. George, and P. Yalla, ‘‘Setting-free fault section identifi-
hardware cost is incurred. cation and fault location method for three-terminal lines,’’ in Proc. 9th Int.
• The proposed method is suitable to calculate the reli- Conf. Power Energy Syst. (ICPES), Dec. 2019, pp. 1–6.
able faulted section and determine the distance to fault [18] A. Hooshyar, M. A. Azzouz, and E. F. El-Saadany, ‘‘Distance protec-
tion of lines emanating from full-scale converter-interfaced renewable
accurately in rugged transmission terrains (for example, energy power plants—Part I: Problem statement,’’ IEEE Trans. Power Del.,
coastal, high/low temperature countries, deserts, etc.) as vol. 30, no. 4, pp. 1770–1780, Aug. 2015.
it does not dependent on the cable parameters and the [19] S. Paladhi and A. K. Pradhan, ‘‘Adaptive distance protection for lines
connecting converter-interfaced renewable plants,’’ IEEE J. Emerg. Sel.
parameters of each section, which vary with change in Topics Power Electron., vol. 9, no. 6, pp. 7088–7098, Dec. 2021.
temperatures. [20] N. George and O. D. Naidu, ‘‘Distance protection issues with renewable
• Moreover, the determined electrical parameters of each power generators and possible solutions,’’ in Proc. 16th Int. Conf. Develop.
Power Syst. Protection, 2022, pp. 1–6.
section can be used for power system monitoring and [21] IEEE Guide for Determining Fault Location on AC Transmission and
protection application settings. Distribution Lines, IEEE Standard C37.114-2014, Jan. 2015.

VOLUME 11, 2023 23855

 K. Likhitha, O. D. Naidu: Setting Free FL for TTHTLs Connected With Conventional and Renewable Resources

[22] M. M. Saha, J. Izykowski, and E. Rosolowski, Fault Location on Power [44] L. Zhang and A. Abur, ‘‘Identifying parameter errors via multiple mea-
Networks (Power Systems). London, U.K.: Springer, 2010. surement scans,’’ IEEE Trans. Power Syst., vol. 28, no. 4, pp. 3916–3923,
[23] E. O. Schweitzer, A. Guzman, M. V. Mynam, V. Skendzic, B. Kasztenny, Nov. 2013.
and S. Marx, ‘‘Locating faults by the traveling waves they launch,’’ in Proc. [45] Y. Liao and M. Kezunovic, ‘‘Online optimal transmission line parameter
67th Annu. Conf. Protective Relay Eng., Mar. 2014, pp. 95–110. estimation for relaying applications,’’ IEEE Trans. Power Del., vol. 24,
[24] H. Shu, X. Liu, and X. Tian, ‘‘Single-ended fault location for hybrid feeders no. 1, pp. 96–102, Jan. 2009.
based on characteristic distribution of traveling wave along a line,’’ IEEE [46] S. S. Mousavi-Seyedi, F. Aminifar, and S. Afsharnia, ‘‘Parameter esti-
Trans. Power Del., vol. 36, no. 1, pp. 339–350, Feb. 2021. mation of multiterminal transmission lines using joint PMU and SCADA
[25] O. Naidu and A. K. Pradhan, ‘‘A traveling wave-based fault location data,’’ IEEE Trans. Power Del., vol. 30, no. 3, pp. 1077–1085, Jun. 2015.
method using unsynchronized current measurements,’’ IEEE Trans. Power [47] Z. Yun, T. Wen, and C. Wang, ‘‘Fault location method for three-terminal
Del., vol. 34, no. 2, pp. 505–513, Apr. 2019. lines in distribution network based on line voltage measured by µMPMU,’’
[26] T. Takagi, Y. Yamakoshi, M. Yamaura, R. Kondow, and T. Matsushima, IEEE Trans. Smart Grid, vol. 12, no. 6, pp. 5095–5112, Nov. 2021.
‘‘Development of a new type fault locator using the one-terminal voltage [48] O. D. Naidu, N. George, and P. Yalla, ‘‘Parameter estimation, selective
and current data,’’ IEEE Power Eng. Rev., vol. PER-2, no. 8, pp. 59–60, auto-reclosing and fault location for three-terminal mixed transmission
Aug. 1982. lines using synchronised data,’’ IET Gener., Transmiss. Distrib., vol. 14,
no. 25, pp. 6049–6060, Dec. 2020.
[27] L. Eriksson, M. M. Saha, and G. D. Rockefeller, ‘‘An accurate fault locator
[49] A. Chowdhury, S. Paladhi, and A. K. Pradhan, ‘‘Adaptive unit protection
with compensation for apparent reactance in the fault resistance resulting
for lines connecting large solar plants using incremental current ratio,’’
from remore-end infeed,’’ IEEE Trans. Power App. Syst., vol. PAS-104,
IEEE Syst. J., vol. 16, no. 2, pp. 3272–3283, Jun. 2022.
no. 2, pp. 423–436, Feb. 1985.
[50] D. Çelik and M. E. Meral, ‘‘Current control based power management
[28] S. Das, S. Santoso, A. Gaikwad, and M. Patel, ‘‘Impedance-based fault
strategy for distributed power generation system,’’ Control Eng. Pract.,
location in transmission networks: Theory and application,’’ IEEE Access,
vol. 82, pp. 72–85, Jan. 2019.
vol. 2, pp. 537–557, 2014.
[51] D. Çelik and M. E. Meral, ‘‘A novel control strategy for grid connected
[29] H. Panahi, R. Zamani, M. Sanaye-Pasand, and H. Mehrjerdi, ‘‘Advances distributed generation system to maximize power delivery capability,’’
in transmission network fault location in modern power systems: Review, Energy, vol. 186, Nov. 2019, Art. no. 115850.
outlook and future works,’’ IEEE Access, vol. 9, pp. 158599–158615, 2021. [52] T. Kauffmann, U. Karaagac, I. Kocar, S. Jensen, E. Farantatos, A. Haddadi,
[30] C. Y. Evrenosoglu and A. Abur, ‘‘Travelling wave based fault location and J. Mahseredjian, ‘‘Short-circuit model for type-IV wind turbine gener-
for teed circuits,’’ IEEE Trans. Power Del., vol. 20, no. 2, pp. 1115–1121, ators with decoupled sequence control,’’ IEEE Trans. Power Del., vol. 34,
Apr. 2005. no. 5, pp. 1998–2007, Oct. 2019.
[31] S. Sawai, A. K. Pradhan, and O. D. Naidu, ‘‘Traveling wave based [53] P. Mishra, A. K. Pradhan, and P. Bajpai, ‘‘Adaptive distance relaying for
fault location of multi-terminal transmission lines,’’ in Proc. IEEE PES distribution lines connecting inverter-interfaced solar PV plant,’’ IEEE
Asia–Pacific Power Energy Eng. Conf. (APPEEC), Nov. 2017, pp. 1–6. Trans. Ind. Electron., vol. 68, no. 3, pp. 2300–2309, Mar. 2021.
[32] A. Ghorbani and H. Mehrjerdi, ‘‘Negative-sequence network based fault
location scheme for double-circuit multi-terminal transmission lines,’’
IEEE Trans. Power Del., vol. 34, no. 3, pp. 1109–1117, Jun. 2019.
[33] D. A. Tziouvaras, ‘‘New multi-ended fault location design for two- or
three-terminal lines,’’ in Proc. 7th Int. Conf. Develop. Power Syst. Pro- KUKKALA LIKHITHA received the B.Tech.
tection (DPSP), 2001, pp. 395–398. degree in electrical and electronics engineering
[34] H. Panahi, M. Sanaye-Pasand, and M. Davarpanah, ‘‘Three-terminal lines from the Visvesvaraya National Institute of Tech-
fault location using two main terminals data in the presence of renewable nology (VNIT), Nagpur, India, in 2018, and the
energy sources,’’ IEEE Trans. Smart Grid, early access, Oct. 25, 2022, doi: M.Tech. degree in smart electric grid from the
10.1109/TSG.2022.3216908. National Institute of Technology (NIT), Warangal,
[35] Y.-H. Lin, C.-W. Liu, and C.-S. Yu, ‘‘A new fault locator for three-terminal India, in 2022. She joined Hitachi Energy India
transmission lines using two-terminal synchronized voltage and current Development Center, Bengaluru, in 2022, as a
phasors,’’ IEEE Trans. Power Del., vol. 17, no. 2, pp. 452–459, Apr. 2002.
Research and Development Engineer. She is
[36] J. Izykowski, E. Rosolowski, M. M. Saha, M. Fulczyk, and P. Balcerek,
involved in the development of relay application
‘‘A fault-location method for application with current differential relays
function library (AFL) related to power system protection and monitoring.
of three-terminal lines,’’ IEEE Trans. Power Del., vol. 22, no. 4,
pp. 2099–2107, Oct. 2007.
She has filed one patent. Her research interests include power system pro-
tection, monitoring applications, and renewable integration.
[37] N. George, O. Naidu, S. Srivastava, and A. V. S. S. R. Sai, ‘‘Fault loca-
tion for multi-terminal tapped lines using synchronized measurements,’’
in Proc. IEEE Recent Adv. Intell. Comput. Syst. (RAICS), Dec. 2018,
pp. 153–158.
[38] T.-C. Lin, J.-Z. Yang, C.-S. Yu, and C.-W. Liu, ‘‘Development of a O. D. NAIDU (Senior Member, IEEE) received
transmission network fault location platform based on cloud computing the Ph.D. degree from the Indian Institute of
and synchrophasors,’’ IEEE Trans. Power Del., vol. 35, no. 1, pp. 84–94,
Technology (IIT) Kharagpur, Kharagpur, India.
Feb. 2020.
From 2009 to 2012, he was a Senior Power Sys-
[39] V. K. Gaur, B. R. Bhalja, and M. Kezunovic, ‘‘Novel fault distance esti-
tem Application Development Engineer with the
mation method for three-terminal transmission line,’’ IEEE Trans. Power
Del., vol. 36, no. 1, pp. 406–417, Feb. 2021. ABB India Development Center, Bengaluru, India.
From 2012 to 2019, he was a Principal Scien-
[40] Y. Liu, B. Wang, X. Zheng, D. Lu, M. Fu, and N. Tai, ‘‘Fault location
algorithm for non-homogeneous transmission lines considering line asym- tist with the ABB Corporate Research Center,
metry,’’ IEEE Trans. Power Del., vol. 35, no. 5, pp. 2425–2437, Oct. 2020. Bengaluru. He is currently a Senior Principal Engi-
[41] G. L. Kusic and D. L. Garrison, ‘‘Measurement of transmission line neer with Hitachi Energy Research and Develop-
parameters from SCADA data,’’ in Proc. IEEE PES Power Syst. Conf. ment Center, Bengaluru. He is the author of more than 50 scientific articles
Expo., New York, NY, USA, 2004, pp. 440–445. and 45 patent applications. He holds 15 granted patents. His research inter-
[42] M. Asprou and E. Kyriakides, ‘‘Identification and estimation of erroneous ests include power system protection, fault location, renewable integration
transmission line parameters using PMU measurements,’’ IEEE Trans. and power system monitoring, and artificial intelligence applications to
Power Del., vol. 32, no. 6, pp. 2510–2519, Dec. 2017. power system protection and monitoring. He was a recipient of the Young
[43] O. D. Naidu, S. Zubic, A. V. S. S. R. Sai, A. N. Praveen, P. Cost, Innovator and Entrepreneur Award from the Indian National Academy of
and H. Eriksson, ‘‘Economical setting-free double-ended fault locator for Engineering (INAE), for his contribution to the field of power system pro-
transmission lines: Experiences from recent pilot installations,’’ IEEE tection and monitoring.
Access, vol. 10, pp. 96805–96820, 2022.

23856 VOLUME 11, 2023