International Journal of Power Electronics and Drive Systems (IJPEDS)

Vol. 14, No. 2, June 2023, pp. 1290~1299

ISSN: 2088-8694, DOI: 10.11591/ijpeds.v14.i2.pp1290-1299  1290

An intelligent overcurrent relay to protect transmission lines

based on artificial neural network

Noha Abed-Al-Bary Al-Jawady1, Mohammed Ahmed Ibrahim1, Laith A. Khalaf1,

Mahmood Natiq Abed2
1
Department of Power Techniques Engineering, Technical College of Engineering, Northern Technical University, Mosul, Iraq
2
Ninavah Electricity Distribution Directorate, General Directorate of North Distribution Electricity, Ministry of Electricity, Mosul, Iraq

Article Info ABSTRACT

Article history: Power systems are susceptible to faults due to system failures or natural
calamities. This could be caused by damage to power system components,
Received Nov 5, 2022 resulting in an interruption of power delivery to clients. Overcurrent relays
Revised Jan 20, 2023 are important relays that protect distribution feeders, transmission lines,
Accepted Feb 2, 2023 transformers, and other components. The intelligent relay can perform both
primary and secondary functions. Line-to-ground (L-G) faults are the most
common occurrence in long transmission lines, posing a serious threat to
Keywords: electrical equipment. This article presents improved fault classification for
transmission line overcurrent protection and highlights the use of artificial
ANN neural network (ANN) techniques to protect transmission lines of 100 km
Fault condition (terco type). An ANN is used to classify the faults. A back propagation
MATLAB/Simulink neural network (BPNN) is used in this case. The neural network has been
Overcurrent relay trained to classify faults in transmission lines for overcurrent protection.
Transmission line protection Various fault conditions are considered. In the event of a fault condition, the
output of a neural network will be a tripping signal. The MATLAB neural
network tool and the Simulink package are used to model the suggested
method.
This is an open access article under the CC BY-SA license.

Corresponding Author:
Mohammed Ahmed Ibrahim
Department of Power Techniques Engineering, Technical College of Engineering
Northern Technical University
95G2+P34, Mosul 41002, Iraq
Email: [email protected]

1. INTRODUCTION
Faults are currently an important issue in the power system network. With rising electricity demand,
the distribution system for electricity is expanding year after year. As a result, protection and maintenance of
power system equipment are crucial for lowering prices and extending the lifetime of dependable and
unbroken components of the power system [1]–[3]. The electricity grid should always be operated safely.
Faults will cause a partial or total system blackout. To safeguard the power system from disruptions,
a protective mechanism is required. To address this issue, a variety of protective relays are available [4].
Overcurrent relays are one type of protection relay that is used to protect power system circuits. By
measuring the current value, an overcurrent relay detects a failure in the power system and separates the
faulty zone from the rest of the system. It can be used to safeguard transmission or distribution feeders,
transformers, bus couplers, and other components as either primary or secondary protection [5]. Fault
analysis can aid in the selection and development of the appropriate protection apparatus. Due to the high
value of current in the three-phase fault compared to other faults, CBs and their specifications should be
introduced as soon as possible [6].

Journal homepage: http://ijpeds.iaescore.com

Int J Pow Elec & Dri Syst ISSN: 2088-8694  1291

Over the last several years, there has been a lot of study done on the applications of artificial neural
networks (ANNs), particularly in the field of the pattern recognition. The ANNs try to replicate the human
brain's processes of generalization and learning [7]. This strategy is focused on distinguishing between
operating states and defects [8]. As an academic study, artificial neural network (ANN) methods are
successfully implemented in many recognition or pattern challenges. ANN approaches are regarded as a
healthy condition for recognition [9].
Each neuron in an artificial neural network performs the task of aggregating external inputs and
providing an output based on the aggregated inputs. Synaptic weight is a numerical strength assigned to a
relationship between two neurons. The development of ANN is divided into two stages: preparation and
testing. ANN training is accomplished by presenting the network with examples referred to as “training sets”
[10]. This study aims at determining the many varieties of rumor faults, which are classified as symmetrical
and asymmetrical faults. To evaluate this circuit and acquire data on the various simulation constraints of
fault types, the MATLAB environment is used.

2. AIM OF THE STUDY

The purpose of the research is to use the ANN technique to identify all kinds of faults, reduce delay
time, and protect transmission lines only during faults in their zone. Because of the good coordination, ANN
output is preferable to computed output. Furthermore, ANN output reduces miscoordination time and
increases relay operational time. This work was utilized in the back-propagation ANN method, and ANN
technology complete an adaptive procedure any defect classifier.

3. LITERATURE REVIEW
Shagila and Rajeswari [10], employed ANN for overcurrent relay protection to identify the fault.
They concluded that the NRDE Board's overcurrent data was trained in ANN and implemented through
coding. When an overcurrent condition is detected in NRDE testing software, a trip command is generated.
Karupiah et al. [11], examined the relay's TMS and PS settings throughout a fault incidence to suggest a new
relay operation time and anticipate the relay's miscoordination time using ANN. They found that the ANN
output is more appropriate than the computed output since it demonstrates strong coordinate. Furthermore,
the ANN output minimizes miscoordination time while increasing relay operational time. Thoeurn et al. [12],
The Bayesian regularization backpropagation neural network (BRBPNN), which is used to represent
nonconventional curves, was investigated. They established that the BRBPNN is a reliable method for
modeling the characteristic curves of overcurrent relays (OCR). It is evident that the errors between real data
and the simulation and real-world use of a prototype are extremely satisfying.

4. THE PROPOSED METHOD

Overcurrent protection is the protection that occurs when the magnitude of the current exceeds the
predicted pick-up value (setting value). A system short circuit usually results in a decrease in circuit
impedance. and hence the failure is accompanied by a significant current. Overcurrent protection safeguards
against overloads [13]. When the input current value exceeds a certain threshold, the overcurrent relay
detects it and sends a signal to the C.B., disconnecting the protected device by opening its contact. When a
fault is detected by the relay, the clause in this case is recognized as “fault pickup”. When a flaw is detected,
the relay can promptly send a trip signal. prompt the overcurrent relay or bid for a predetermined amount of
time before issuing a trip signal [14]–[17]. The “relay operational time” is the time it takes the relay to
calculate it, and it is calculated using the microprocessor's protection algorithm [18], [19]. The logical form
of an overcurrent relay is shown in Figure 1.

Figure 1. Overcurrent relay logical representation

An intelligent overcurrent relay to protect transmission lines based … (Noha Abed-Al-Bary Al-Jawady)
1292  ISSN: 2088-8694

Overcurrent relays are categorized into three classes based on their operation time:
− Instantaneous overcurrent relays: When a problem is detected, these relays immediately send a trip
instruction to the breaker (because the value of the input current is greater than the present value). They
have no deliberate timing lag. They are typically implemented near the source, where the fault's current is
relatively large and a minor delay in relay functioning could result in catastrophic equipment damage.
− Definite time overcurrent relay (DTOR): This sort of overcurrent relay serves as a backup protection
device. When there is a fault in the line and the distance relay is not detected and does not send a signal to
C.B., after a predetermined delay, the relay sends a signal to C.B. In this situation, the relay has been
delayed for a duration significantly longer than the distance relay's normal operating time as well as the
operating time of C.B.
− Inverse definite minimum time overcurrent relay: A relay's operating time is inversely proportional to the
fault current and vice versa. When the fault current is very high, the operating time is shortened. It may be
rated for a wide range of fault currents and operation times [1], [20]. The flowchart of the overcurrent
relay protection for transmission lines is depicted in Figure 2.

Figure 2. Flowchart of ANN structure for transmission line protection

The input data is the load currents, and the pickup current is set to 1.2 A. The fundamental
overcurrent prevention algorithm compares the measured current value to a preset value. If the value of the
input current is greater than the present value, as confirmed via simulation, it is supplied in the ANN
functional fitting training in MATLAB to find an improved ideal relay operation time. As a result, the ANN

Int J Pow Elec & Dri Syst, Vol. 14, No. 2, June 2023: 1290-1299
Int J Pow Elec & Dri Syst ISSN: 2088-8694  1293

values acquired can be utilized to generate a new time setting for the relay operation as well as anticipate the
more likely timing of the relay operation. When an overcurrent condition is detected, the protection algorithm
transmits a signal to the breaker, which in turn cuts off the protected circuit.

5. MATERIALS AND METHOD

The data for this study was obtained from a Swedish company for a three-phase model of a 100-
kilometer-long overhead electricity transmission line (Terco Company). The Terco-prototype of a
transmission line, as illustrated in Figure 3, was used in this investigation; Table 1 shows the data of the line
used. The multipurpose relay model is evaluated under various fault conditions. A Terco (MV2221)
transmission line was used in the scheme of protection's design, and the MATLAB/Simulink environment
was used for the implementation.

Figure 3. Terco-prototype of transmission line (MV2221 line model)

Table 1. The Specifications of Terco-prototype of transmission line

Parameters Terco company
Lounge of transmission lines. 100 Km
Voltage 230 kV
Ability 110 MVA
Resistance 2.20 ohm
Inductance 25 mH
Capacitance 4 uF
Dimensions 410×245×160 mm
Weight 10 kg

6. RESEARCH METHOD
MATLAB provides an appealing environment with hundreds of dependable and efficient built-in
algorithms. These functions aid in the solution of a wide range of mathematical problems, including linear
systems, matrix algebra, nonlinear systems, differential equations, and many other sorts of electrical
machine-related technical solutions. Figure 4 shows the modeling circuit for over-current relay protection.

Figure 4. Over current relay protection circuit modeling

An intelligent overcurrent relay to protect transmission lines based … (Noha Abed-Al-Bary Al-Jawady)
1294  ISSN: 2088-8694

7. ANN TECHNIQUES
The ANN is seen as a machine that functions similarly to the human brain. It possesses learning,
capacity, and generalization qualities [21]. It may also be characterized as a group of artificial neurons that
rely on a set of separate mathematics models for information processing, which are based on a relationship
method of computing [22]. The most significant component of this method is its ability to grasp and
understand the model's behavior through prolonged data training, resulting in satisfactory and desirable
outputs. As a result, the ANNs algorithm is used to identify the different types of faults [23], [24]. ANNs
methods were trained with a range of parameters, and the suggested relay-based ANNs were trained with
“nntool”. The number of output neurons was decided by the goal number, the number of neurons in the input
layer, on the other hand, was determined by the number of training samples. Finally, the number of hidden
neurons within the hidden layer was calculated through trial and error [25]. The internal structure of the
proposed ANN is illustrated in Figure 5.
Back-propagation is the most widely used and effective method for generating multilayer feed-
forward networks. Backpropagation learning is divided into two stages: forward and backward. During the
forward phase, the input signals pass across the network layer, finally creating a response at the network's
output. When an actual answer differs from the anticipated (target) response, error signals are generated and
sent backward through the network. The network's free parameters are modified during the backward phase
of operation to reduce the sum of squared errors. Back-propagation learning has been used successfully to
solve a wide range of challenging problems [26]. The output layer was connected to the hidden layers, as
seen in Figure 6. Figures 7 and 8 illustrate the performance and state plot of the trained ANN. Figure 9
illustrates the regression performance plot.

Figure 5. Internal structure of the proposed ANN Figure 6. ANN's internal architecture

Figure 7. The effectiveness plot of the trained ANN

Int J Pow Elec & Dri Syst, Vol. 14, No. 2, June 2023: 1290-1299
Int J Pow Elec & Dri Syst ISSN: 2088-8694  1295

Figure 8. State plot of the trained ANN

Figure 9. Regression performance plot

8. DISCUSS THE RESULT

8.1. Normal operation
In the power system's normal operation, if the current value is less than the reference current (setting
current), the trip signal is not sent to a circuit breaker (C.B.). Figure 10 represents the output signal of the
intelligent relay without faults, where a value of one means there is no fault and a value of zero means there
is a fault. Figure 11 shows the current in the transmission line with no faults.

8.2. Fault situation

In the power system's fault situation, if there is a fault in one of the lines with ground, the fault
current becomes very high compared to the reference current (setting current). In this case, the trip signal is
sent to the C.B. after 4.4 ms, which is very fast. Figure 12 represents the output signal of the intelligent relay
with faults, where a value of one means there is no fault and a value of zero means there is a fault. Figure 13
depicts the transmission line current with one line faulted to ground. Figure 14 depicts the transmission line
current with tow line faults to ground. Figure 15 illustrate the transmission line current with faults on three
lines to ground.
An intelligent overcurrent relay to protect transmission lines based … (Noha Abed-Al-Bary Al-Jawady)
1296  ISSN: 2088-8694

Figure 10. Output signal of the intelligent relay without faults

Figure 11. Current of the transmission line with no faults

Figure 12. Output signal of the intelligent relay without faults

Int J Pow Elec & Dri Syst, Vol. 14, No. 2, June 2023: 1290-1299
Int J Pow Elec & Dri Syst ISSN: 2088-8694  1297

Figure 13. Transmission line current with faults on one line to ground

Figure 14. Transmission line current with faults on two lines to ground

Figure 15. Transmission line current with faults on three lines to ground

An intelligent overcurrent relay to protect transmission lines based … (Noha Abed-Al-Bary Al-Jawady)
1298  ISSN: 2088-8694

9. CONCLUSION
In the MATLAB environment, the ANN structure represented the overcurrent relay properties. A
new overcurrent protection algorithm for transmission line faults was used to simulate a 100-km transmission
line (Terco type) in MATLAB/Simulink with actual parameters. Furthermore, transmission line currents with
defects on lines 1, 2, and 3 to ground were investigated. The results indicate that when no faults occur and the
current is normal, the current value is less than the reference current, and the trip signal is not transmitted to a
circuit breaker (C.B.). The results show that if a fault occurs in one, two, or three of the grounding lines, the
fault current becomes quite high in comparison to the reference current (setting current). After 4.4 ms, the
trip signal is forwarded to a circuit breaker (C.B.). This very short time gives great importance and reliability
to protecting the electrical equipment in the electric power system. According to the simulation findings, the
proposed overcurrent relay is a good solution. The dependability, security, and speed of operation of the
ANN-based overcurrent relay appears to be promising for power system protection. The results suggest that
the proposed technique yields considerable outcomes. Furthermore, ANN output reduces miscoordination
time and speeds up relay operation time, while the time between the open circuit breaker and the trip signal is
very short. That means the intelligent relay used gave good results.

ACKNOWLEDGEMENTS
The authors would like to thank Northern Technical University/Technical College of
Engineering/Mosul for providing us with a simulation package that enabled us to finish our research.

REFERENCES
[1] P. Maji and G. Ghosh, “Designing over-current relay logic in MATLAB,” International Journal of Scientific & Engineering
Research, vol. 8, no. 3, pp. 40–43, 2017, [Online]. Available: http://www.ijser.org.
[2] A. H. Al-Rifaie, M. A. Ibrahim, and H. A. T. Abdullah, “Design and implementation of differential relay based on graphical user
interface,” Przeglad Elektrotechniczny, vol. 98, no. 7, pp. 188–192, 2022, doi: 10.15199/48.2022.07.31.
[3] B. M. Salih, M. A. Ibrahim, and A. N. Hamoodi, “Differential relay protection for prototype transformer,” Przeglad
Elektrotechniczny, vol. 97, no. 6, pp. 158–162, 2021, doi: 10.15199/48.2021.06.30.
[4] M. A. Ibrahim, B. M. Salih, and M. N. Abd, “Protection transformer and transmission line in power system based on MATLAB
simulink,” Przeglad Elektrotechniczny, vol. 97, no. 10, pp. 17–21, 2021, doi: 10.15199/48.2021.10.04.
[5] N. H. Hussin et al., “Modeling and simulation of inverse time overcurrent relay using Matlab/Simulink,” Proceedings - 2016
IEEE International Conference on Automatic Control and Intelligent Systems, I2CACIS 2016, pp. 40–44, 2017, doi:
10.1109/I2CACIS.2016.7885286.
[6] A. N. Hamoodi, M. A. Ibrahim, and B. M. Salih, “An intelligent differential protection of power transformer based on artificial
neural network,” Bulletin of Electrical Engineering and Informatics, vol. 11, no. 1, pp. 93–102, 2022, doi:
10.11591/eei.v11i1.3547.
[7] D. N. Vishwakarma and Z. Moravej, “Ann based directional overcurrent relay,” Proceedings of the IEEE Power Engineering
Society Transmission and Distribution Conference, vol. 1, pp. 59–64, 2001, doi: 10.1109/TDC.2001.971209.
[8] H. Balaga, N. Gupta, and D. N. Vishwakarma, “GA trained parallel hidden layered ANN based differential protection of three
phase power transformer,” International Journal of Electrical Power and Energy Systems, vol. 67, pp. 286–297, 2015, doi:
10.1016/j.ijepes.2014.11.028.
[9] D. K. Singh and S. Gupta, “Optimal coordination of directional overcurrent relays: A genetic algorithm approach,” 2012 IEEE
Students’ Conference on Electrical, Electronics and Computer Science: Innovation for Humanity, SCEECS 2012, 2012, doi:
10.1109/SCEECS.2012.6184808.
[10] S. Shagila and R. Rajeswari, “Implementation Of AI based relaying scheme for overcurrent protection using numerical relay
development environment board,” International Research Journal on Advanced Science Hub, vol. 2, no. 6, pp. 98–104, 2020, doi:
10.47392/irjash.2020.44.
[11] S. Karupiah, M. H. Hussain, I. Musirin, and S. R. A. Rahim, “Prediction of overcurrent relay miscoordination time using urtificial
neural network,” Indonesian Journal of Electrical Engineering and Computer Science, vol. 14, no. 1, pp. 319–326, 2019, doi:
10.11591/ijeecs.v14.i1.pp319-326.
[12] M. Thoeurn, A. Priyadi, A. Tjahjono, and M. H. Purnomo, “Overcurrent relay modeling using artificial neural network,” 2017
International Electrical Engineering Congress, iEECON 2017, 2017, doi: 10.1109/IEECON.2017.8075794.
[13] A. Akhikpemelo and M. J. E. Evbogbai, “Overcurrent relays coordination using MATLAB model,” vol. 7, no. 2, p. 1, 2018.
[14] M. A. Ibrahim, W. K. Ibrahim, and A. N. Hamoodi, “Design and implementation of overcurrent relay to protect the transmission
line,” International Journal of Engineering Research and Technology, vol. 13, no. 11, pp. 3783–3789, 2020, doi:
10.37624/ijert/13.11.2020.3783-3789.
[15] D. Tzelepis et al., “Impact of synchronous condensers on transmission line protection in scenarios with high penetration of
renewable energy sources,” IET Conference Publications, vol. 2020, no. CP771, 2020, doi: 10.1049/cp.2020.0095.
[16] B. K. Sugumar and S. Balaraman, “Implementation of differential algorithm for busbar protection,” International Research
Journal of Multidisciplinary Technovation, pp. 150–157, 2019, doi: 10.34256/irjmt19220.
[17] M. S. Almas, R. Leelaruji, and L. Vanfretti, “Over-current relay model implementation for real time simulation & Hardware-in-
the-Loop (HIL) validation,” IECON Proceedings (Industrial Electronics Conference), pp. 4789–4796, 2012, doi:
10.1109/IECON.2012.6389585.
[18] Y. Yang, L. Zhu, F. Li, Z. Wu, and C. Cai, “Research on the power line three-stage over-current protection simulation,” 2016,
doi: 10.2991/icadme-16.2016.3.
[19] P. V Warbhe and S. K. Malode, “Over-current relay and circuit breaker designing and testing,” IJARIIE, vol. 2, no. 3, pp. 1432–
1439, 2017.
[20] A. Musirikare, M. Pujiantara, A. Tjahjono, and M. H. Purnomo, “ANN-Based modeling of directional overcurrent relay

Int J Pow Elec & Dri Syst, Vol. 14, No. 2, June 2023: 1290-1299
Int J Pow Elec & Dri Syst ISSN: 2088-8694  1299

characteristics applied in radial distribution system with distributed generations,” Proceedings of 2018 10th International
Conference on Information Technology and Electrical Engineering: Smart Technology for Better Society, ICITEE 2018, pp. 52–
57, 2018, doi: 10.1109/ICITEED.2018.8534834.
[21] E. B. M. Tayeb and O. A. A. A. Rhim, “Transmission line faults detection, classification and location using artificial neural
network,” Proceedings of the 2011 International Conference and Utility Exhibition on Power and Energy Systems: Issues and
Prospects for Asia, ICUE 2011, 2012, doi: 10.1109/ICUEPES.2011.6497761.
[22] A. P. Vaidya and P. A. Venikar, “ANN based distance protection of long transmission lines by considering the effect of fault
resistance,” IEEE-International Conference on Advances in Engineering, Science and Management, ICAESM-2012, pp. 590–594, 2012.
[23] M. S. Islam and M. M. Kabir, “ANN based discrimination of inrush and fault currents in three phase power transformer using
statistical approaches,” 2019 4th International Conference on Electrical Information and Communication Technology, EICT
2019, 2019, doi: 10.1109/EICT48899.2019.9068766.
[24] Shah and M. S. B. Parmar, “Transformer protection using artificial neural network,” International Journal of Novel Research and
Development, vol. 2, pp. 2456–4184, 2017, [Online]. Available: http://www.ijnrd.org/papers/IJNRD1705023.pdf.
[25] E. Ali, A. Helal, H. Desouki, K. Shebl, S. Abdelkader, and O. P. Malik, “Power transformer differential protection using current
and voltage ratios,” Electric Power Systems Research, vol. 154, pp. 140–150, 2018, doi: 10.1016/j.epsr.2017.08.026.
[26] M. Bakkar, S. Bogarra, F. Córcoles, and J. Iglesias, “Overcurrent protection based on ANNs for smart distribution networks with grid-
connected VSIs,” IET Generation, Transmission and Distribution, vol. 15, no. 7, pp. 1159–1174, 2021, doi: 10.1049/gtd2.12093.

BIOGRAPHIES OF AUTHORS

Noha Abed-Al-Bary Al-Jawady was born in Mosul, Iraq in 1972, She received
the M.Sc degrees in electrical engineering from the college engineering electrical engineering
department in the University of Mosul, Iraq, in 1999. After graduation, she joined the
Northern Technical University, Mosul, Iraq, Technical engineering college of Mosul, as a
Lecturer. She has many research in field of protection power system, Renewable energy and
A.C. Machine. She can be contacted at email: [email protected].

Mohammed Ahmed Ibrahim is an engineer Department of Electrical Power

Technologies Engineering. He received bachelor's degree from the Northern Technical
University 2005 in Electrical Engineering, master's degree from the Northern Technical
University in (2011) in Electrical Power Technologies Engineering. Currently, he is working
as a lecturer in the Electrical Power Engineering specializing, at Northern Technical
University, Technical College of Engineering-Mosul-Iraq. He can be contacted at email:
[email protected].

Laith A. Khalaf was born in Mosul, Iraq in 1972, he received the M.Sc degrees
in electrical engineering from the electrical engineering department in the University of
Technology, Iraq, in 2001. After graduation, he joined the Northern Technical University,
Mosul, Iraq, Technical engineering college of Mosul, as a Lecturer. he has many research in
field of power electronics, speed control, power system and electrical machines. He can be
contacted at email: [email protected].

Mahmood Natiq Abed is an engineer Department of high voltage underground

cables test and detection their faults. He received bachelor's degree from the Northern
Technical University 2005 in Electrical Engineering, master's degree from the University of
Turkish Aeronautical Association in (2017) in electrical and electronic engineering. Currently,
he is working as an employee in Ninavah Electricity Distribution Directorate at General
Directorate Electricity for North Mosul, Iraq. He can be contacted at email:
[email protected].

An intelligent overcurrent relay to protect transmission lines based … (Noha Abed-Al-Bary Al-Jawady)