Audit-based Power Surge Detection using Federated Learning in Smart
Transmission Lines
M. MOHAMMED THAHA1, ROSINI NAWANG MUSTAPEN2, RAFIKULLAH DERAMAN3,
SHANMUGAM DURAIRAJ4, RAJENDRAKUMAR RAMADASS4
1B.S.Abdur Rahman Crescent Institute of Science and Technology,
GST Road, Vandalur, Chennai - 600 048, Tamilnadu,
INDIA
2School of Technology Management and Logistics,
Universiti Utara Malaysia,
Kedah,
MALAYSIA
3Project and Facilities Management Research Group,
Faculty of Civil Engineering and Built Environment,
Universiti Tun Hussein Onn Malaysia,
Batu Pahat,
MALAYSIA
4College of Engineering and Technology, Engineering Department,
University of Technology and Applied Sciences,
Shinas,
SULTANATE OF OMAN
Abstract: - Smart transmission lines are designed to improve the optimal distribution irrespective of the surge
due to peak utilization and generation. Therefore distribution audits are mandatory for identifying power surges
in these transmission lines. This article, therefore, proposes a Power Surge Detection using the Transmission
Audit (PSD-TA) scheme. The proposed scheme houses federated learning for identifying surges due to
generation or utilization between distribution points. Based on the detection, the regulation or transmission
allocation for the distinct surges is recommended by the learning for reducing failures. Therefore the previous
audit from the surge is used for identifying similar failures by training the learning paradigm. This scheme,
therefore, improves the distribution rate and meets the utilization demands of the users.
Key-Words: - Audit Process, Federated Learning, Power Surge, Smart Transmission, PSD-TA, failure,
distribution rate, peak utilization.
Received: November 27, 2022. Revised: November 28, 2023. Accepted: December 13, 2023. Published: December 31, 2023.
1 Introduction
A power surge is an abnormally high voltage that
occurs for a short period. The power surge is a
transient wave of voltage in an electrical circuit.
Power surge detection is a crucial task to perform in
every smart transmission line, [1]. An effective
identification method is used to identify the power
surges in transmission lines. The identification
method optimizes the inputs from storage and
detects the abnormal voltage or shortage in the
circuit, [2]. A cross-entropy optimization algorithm
is used here to identify the multiple surges that are
presented in the lines. The optimization algorithm
reduces the problems which are occurred during the
surge detection process, [3]. The identification
method improves the overall outage detection
performance range and enhances the feasibility
range of smart transmission lines. Long short-term
memory (LSTM) algorithm-based power surge
detection method is also used to detect the activities
of transmission lines. The LSTM algorithm
classifies the exact types of power surges based on
the condition and outage ratio of the circuits. The
LSTM algorithm uses a fault line dataset to detect
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2023.18.37
M. Mohammed Thaha, Rosini Nawang Mustapen,
Rafikullah Deraman, Shanmugam Durairaj,
Rajendrakumar Ramadass
E-ISSN: 2224-350X
364
Volume 18, 2023
the impact of outages in smart transmission lines.
The LSTM-based method increases the accuracy of
the power surge detection process, [4], [5].
An audit-based method and techniques are used
for the power surge rectification process in smart
transmission lines. The audit-based method
identifies the exact power quality (PQ) and issues
that are presented in the lines, [6]. The audit-based
method creates a major impact in improving the
performance range of transmission lines, [7]. A
machine learning (ML) based audit technique is
used to evaluate the actual power surge ratio in the
lines. The ML-based technique is used as an
adaptive surge detection technique that provides
feasible data for the rectification process, [8]. The
audit technique produces the key characteristics of
power surge which create an active reason to rectify
the power supply in transmission lines. The ML-
based technique improves the reliability range of the
lines, [9]. A hybrid rectification model is also used
to eliminate the power surges in transmission lines.
The hybrid model identifies the parameters and
variables that are necessary for the rectification
process. The hybrid model reduces the latency and
energy consumption level in the computation
process. The important values are audited from the
storage which enhances the effective range of the
systems, [10].
The Proposed method is used to overcome the
above issues and to detect the surge with
distribution rate and failure detection. The
remaining part of this paper is organized as follows:
Section 2 discusses the related works, Section 3
proposes the Audit Scheme for the smart
transmission line depending on the surge detection;
Section 4 describes the results and discussion;
Section 5 presents the Distribution Rate; Section 6
evaluates the results of surge detection; Section 7
finds the failures using the federated learning
techniques and finally Section 8 Concludes this
work.
2 Related Works
The study, [10], proposed a hybrid finite-difference
time-domain and partial element equivalent circuit
(FDTD-PEEC) method for lightning power surge
analysis. The proposed method is mostly used in
transmission lines that detect lightning surges.
PEEC validates the channels that contain surges by
analyzing the circuits. FDTD identifies the
weakened channels based on their coupling. The
proposed method reduces the latency in the data
transmission process which improves the
performance level of the systems.
The study, [11], developed a new multi-criteria
method based on an optimization algorithm for
transmission line surge arrestors (TLSA) detection.
The main aim of the method is to provide optimal
solutions to solve surge arrestors in transmission
lines. A lightning performance calculation technique
is used in the method which analyzes the exact
effectiveness of power surge arrestors. The
developed method increases the accuracy of TLSA
detection which enhances the feasibility range of
transmission lines.
The study, [12], introduced an intelligence
control strategy based on a radial basis function
neural network (RBF-NN). The introduced strategy
is used as an anti-surge control strategy in
transmission lines. The RBF-NN controls the
derivation that occurs during the transmission
process. The introduced strategy improves the self-
efficacy and flexibility range of the systems.
Experimental results show that the introduced
strategy increases the adaptability and efficient ratio
of transmission lines.
The study, [13], proposed a steady-state voltage-
control method using half-wavelength transmission
lines. The actual goal is to reduce the unique voltage
range in transmission lines. The proposed method
identifies the active voltage-control level of the lines
which produces optimal features for controlling the
process. The proposed method is used as a
secondary voltage-control method which enhances
the performance and realisability range of half-
wavelength transmission lines.
The study, [14], designed an optimization model
for inherent flexibility in transmission lines. A
stochastic unit commitment framework is
implemented in the model to analyze the exact
weather conditions of the lines. The analyzed data
produce relevant information for decision-making
which increases the flexibility and robustness range
of transmission lines. The optimization model
reduces the latency in the computation process. The
designed model improves the energy-efficiency
level of renewable generation systems.
The study, [15], proposed a transient recovery
voltage (TRV) analysis method for high-voltage
transmission lines. The proposed TRV analyses the
voltage during a short circuit (SC) which predicts
the breaks in lines. The proposed analysis method
identifies the circuit breakers and important
characteristics of transmission lines. The proposed
method increases the clearing ability of lines which
improves the feasibility range of high-voltage
transmission lines.
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2023.18.37
M. Mohammed Thaha, Rosini Nawang Mustapen,
Rafikullah Deraman, Shanmugam Durairaj,
Rajendrakumar Ramadass
E-ISSN: 2224-350X
365
Volume 18, 2023
3 Proposed Audit Scheme
Smart power transmission lines are a technological
advancement aimed at improving the efficacious
distribution of power, regardless of the surges
caused by peak utilization or low utilization.
Fig. 1: Proposed PSD-TA Scheme
To determine this, organizing the dispensing of
the audits becomes significant in identifying the
power surges within these transmission lines. The
proposed scheme is illustrated in Figure 1.
The proposed Architecture of the PSD-TA
Scheme explains the workflow of the proposed
system. The smart power transmission has been
partitioned into several distribution points and each
distribution point is used to detect the surge in
transmission lines the proposed PSD-TA-based
federated learning detects the error failures in the
learning system and finally provides the modified
transmission.
By establishing the PSD-TA scheme, the
distribution rate of power is enhanced significantly.
This scheme permits the power transmission lines to
match up with the growing demands by effectively
managing and distributing the transmission
resources depending on the surge detection. This
scheme not only enhances the power distribution
process but also improves the overall responsibility
and security of the transmission lines. Furthermore,
the PSD-TA scheme is an instigating approach to
the detection of surges in smart transmission lines.
By utilizing the abilities of federated learning and
historical audit data, this scheme effectively
determines and identifies the power surges,
consequently enhancing the distribution rate and
increasing the demands of power users. The process
is in line with the suggestion by the previous
studies, [16], [17].
The smart transmission lines are analyzed for the
further extraction of the distribution points. It is
used to determine the data from the transmission
lines where the power is generated accurately. By
analyzing the data from the transmission lines, the
distribution points are determined and then it is used
in the identification of surges in the transmission
lines. It is also utilized for the extraction of the
stability and reliability of the transmission lines. The
process of analyzing the smart transmission lines is
explained by the following equation given below:
(1)
Where is denoted as the analyzing process, is
represented as the data presented in the transmission
lines which is used in the further process. Now the
distribution lines are extracted from the data
collected in the smart transmission lines. These
distribution points help detect the surges that occur
in the transmission lines by using the data collected
from them. These distribution points are helpful in
the management of the load balancing, and then the
effective distribution of the power to the customers,
[18]. It also enables better planning in the process
and effective optimization of the power distribution.
The process of extracting the distribution points
from the smart power transmission lines is explained
by the following equation given below:
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2023.18.37
M. Mohammed Thaha, Rosini Nawang Mustapen,
Rafikullah Deraman, Shanmugam Durairaj,
Rajendrakumar Ramadass
E-ISSN: 2224-350X
366
Volume 18, 2023
(2)
Where is the extracted distribution points, is
denoted as the management of power operation.
Now the surge that occurred in the transmission
lines is determined by using the federated learning
technique. The distribution audits are important for
determining the power surges in these transmission
lines. The surges are identified during the power
generation or utilization between the distribution
points of smart transmission lines, [19]. The
federated learning analyses the transmission lines
and then the data collected from it to detect surges
during the peak utilization or generation of power.
The process of detecting the surge between the
distribution points is explained by the following
equation given below:
(3)
Where is denoted as the surge that occurred
between the distribution points, is represented as
the distribution audits, and is represented as the
peak utilization of the power in the transmission
lines. Now the power generation and power
utilization are determined by using the federated
learning technique. The surge detection process is
illustrated in Figure 2.
Fig. 2: Surge Detection Process
The distribution points are monitored for their
power demand for which is calibrated. Depending
on generated/ utilized the drop/ peak is estimated
across various for precise. In this case, the
between successive is detected for estimation
(Figure 2). This process enables the precise
determination and optimization of power generation
levels depending on the usage of the customers. By
using the federated learning technique, it enhances
the generation and utilization of power capacity and
improves the efficaciousness of the power supply.
After identifying the power utilization and power
generation, the failures analyzing process arise. The
process of determining the power utilization and
power utilization from the surge occurred by using
federated learning is explained by the following
equations given below:
(4)
(5)
Where is represented as the quantity of power
generation, is denoted as the quantity of power
utilization, and is represented as the efficiency of
the power supply. After this determination process,
the failures in this process are analyzed and then
some of the steps are taken to reduce the failures.
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2023.18.37
M. Mohammed Thaha, Rosini Nawang Mustapen,
Rafikullah Deraman, Shanmugam Durairaj,
Rajendrakumar Ramadass
E-ISSN: 2224-350X
367
Volume 18, 2023
Fig. 3: Learning Process Representations
To reduce the failures that are caused by the
occurrence of surges, the learning paradigm
influences the previous surge detection to conform
to the regulatory criteria and transmission
allocations, [20]. By training the federated learning
model with the audit data from the past surge, it
becomes efficient to determine the patterns and
issues associated with the failures. The process of
determining and mitigating the failures after the
federated learning process is explained by the
following equation given below:
(6)
Where is denoted as the failures determined,
is represented as the mitigating operation of the
failures. Now the modified transmissions arise by
reduced failures and precise data collection from the
smart power transmission lines. The learning
process is illustrated in Figure 3.
The is differentiated for two intervals such as
and for such that the identifies . If for
and are not the same then occurs. This is
validated based on the available at any.
Therefore the multiple are modified for balancing
distribution based on and logs at. This is
recurrent until without (Figure 3). The
modification is done to enhance the distribution rate
and then to mitigate the failures in the upcoming
process. By considering the previous surge audits,
the modifications in the smart transmission process
are arising for the reduction of the failures in the
upcoming procedures. The modified transmission
process is explained by the following equation given
below:
(7)
Where is denoted as the modified
transmissions. This proposed method helps in
detecting the distribution points from the
transmission lines for further federated learning
procedures. After this, the surges are detected from
the peak utilization or generation between the
distribution points. The failures are identified after
the learning process and then the mitigation process
is done to reduce the failures. This process enhances
the distribution rate and helps in meeting the
utilization demands of the users. Based on the above
discussion, log analysis for and are analyzed
in Figure 4.
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2023.18.37
M. Mohammed Thaha, Rosini Nawang Mustapen,
Rafikullah Deraman, Shanmugam Durairaj,
Rajendrakumar Ramadass
E-ISSN: 2224-350X
368
Volume 18, 2023
Fig. 4: Analysis for and
The is performed for different occurrences
of. Based on the learning instances
for the is identified at the first
distribution. In the consecutive transmission
process, the based analysis is performed for
mitigating. For such a process the or are
performed if (or) analysis is performed
(Figure 4). When compared to other existing works,
the proposed work detects the surge detection with
the distribution rate, surge detection, and failure rate
using proposed PSD-TA scheme of the smart power
transmission lines of the analysis for and
through federated learning.
4 Results and Discussion
This section presents a comparative discussion of
distribution rate, surge detection, and failure rate by
varying the logs/ distribution point and the number
of distribution points. The data from, [21], is utilized
for analyzing the aforementioned metrics. The
existing SSVCM, [13], and SAASIC, [12], methods
are utilized in this comparative analysis.
5 Distribution Rate
Fig. 5: Distribution Rate
The distribution rate is efficacious in this method
by using the federated learning technique. The data
is collected from the transmission lines and then the
distribution points are determined from it. It
measures the efficacy of the process and detects
how accurately the distribution process is done by
satisfying the utilization demands of the users. By
aiding the distribution points from the transmission
lines, the distribution rate is enhanced and thus it
helps in detecting the surge between the distribution
points (Figure 5).
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2023.18.37
M. Mohammed Thaha, Rosini Nawang Mustapen,
Rafikullah Deraman, Shanmugam Durairaj,
Rajendrakumar Ramadass
E-ISSN: 2224-350X
369
Volume 18, 2023
6 Surge Detection
Fig. 6: Surge Detection
The detection of the surge is better in this
process by using the federated learning technique
and thus the power generation and then the power
utilization is also obtained from the surge detection.
When the peak utilization or generation of the
power occurs in the transmission lines, then the
surge is identified. This learning technique is trained
by the acquired paradigm for the detection of the
surge occurrence in the smart power transmission
lines and then further steps are taken to mitigate the
surges (Figure 6).
The failures are lesser in this process after
mitigating the surges between the distribution lines
of the smart transmission lines. To remit the failures
that are caused by the circumstances of surges, the
learning paradigm impacts the previous surge
detection to confirm the regulatory criteria and
transmission allocations. By training the federated
learning model with the audit data from the past
surge, it becomes efficient to determine the patterns
and issues associated with the failures (Figure 7).
Federated learning has been integrated into the
suggested method to detect surges caused by
generation or utilization between distribution sites.
Learning suggests regulating or allocating
transmission for individual surges using detection to
minimize failures. As a result, the framework for
learning is trained using the prior examination from
the surge to find comparable errors.
7 Failures
Fig. 7: Failures
8 Conclusion
In this article, a novel scheme called Power Surge
Detection using Transmission Audit (PSD-TA) is
proposed for the incorporation of the federated
learning techniques to determine the surges
occurring due to production or utilization variance
between the power distribution points. This
proposed method also helps in detecting the surges
within the transmission lines and it classifies the
different types of surges accurately from the data
collected from the distribution points. This validates
the system to suggest precise regulatory or
transmission allocation measures to reduce the
failures caused by these surges. The proposed
scheme is active in improving surge detection by
12.88% and reducing the failure rate by 8.5% for the
varying distribution points.
Federated learning is a novel method that
supports distributed learning for on-device
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2023.18.37
M. Mohammed Thaha, Rosini Nawang Mustapen,
Rafikullah Deraman, Shanmugam Durairaj,
Rajendrakumar Ramadass
E-ISSN: 2224-350X
370
Volume 18, 2023
algorithms. The goal of FL was to expand the
advantages of federated learning to contexts that
have intelligent transmission lines. We present a
thorough analysis of power surge detection in the
FL circumstances, including its advantages,
challenges, and results. With the analysis and
findings on distribution rate, surge detection, and
failure rate, the developers are trying to present
innovative knowledge and raise the community’s
emphasis on building safe FL ecosystems, suitable
for general application. With the future directions
sections, this paper describes the fields in FL that
deserve in-depth study and research. Since FL
remains a very recent technology in this sector,
more research is necessary to determine which
enhanced top-ups are appropriate for the various FL
environment types.
References:
[1] Zhu, Q., Zhang, Y., Ma, Y., Zhou, F., Yin, H.,
Zhang, H., & Qiu, T. (2023). Experimental
and Analytical Studies on Lightning Surge
Response of the Quadruple-Circuit
Transmission Line. IEEE Access, 11, 6879-
6886.
[2] Munir, A., Abdul-Malek, Z., Sultana, U., &
Arshad, R. N. (2022). A novel technique for
condition monitoring of metal oxide surge
arrester based on fifth harmonic resistive
current. Electric Power Systems Research,
202, 107576.
[3] Visacro, S., Silveira, F. H., Pereira, B., &
Gomes, R. M. (2020). Constraints on the use
of surge arresters for improving the
backflashover rate of transmission lines.
Electric Power Systems Research, 180,
106064.
[4] Baskar, S., Dhote, S., Dhote, T., Akila, D., &
Arunprathap, S. (2022). Surge detection for
smart grid power distribution using a
regression-based signal processing model.
Computers and Electrical Engineering, 104,
108424.
[5] Misrikhanov, M. S., & Mirzaabdullaev, A. O.
(2020). On application features of nonlinear
surge suppressors on overhead power
transmission lines. Power Technology and
Engineering, 54, 570-574.
[6] Deligant, M., Pereira, M., Laleg-Kirati, T. M.,
Bakir, F., & Khelladi, S. (2022). Toward a
detection approach of surge precursors using a
semi-classical signal analysis method. The
European Physical Journal Plus, 137(6), 1-
14.
[7] Kanatani, K., Matsuura, S., Fujita, H., &
Michishita, K. (2022). Influence of
configuration of power distribution lines on
failure probability of surge arresters. Electric
Power Systems Research, 213, 108769.
[8] Moro, A. F., Ortega, J. S., & Tavares, M. C.
(2022). Performance evaluation of power
differential protection applied to half-
wavelength transmission lines. Electric Power
Systems Research, 209, 107998.
[9] Cao, D., Yuan, C., Wang, D., & Huang, X.
(2022). Transition from Unsteady Flow
Inception to Rotating Stall and Surge in a
Transonic Compressor. Journal of Thermal
Science, 31(1), 120-129.
[10] Cao, J., Du, Y., Ding, Y., Li, B., Qi, R.,
Zhang, Y., & Li, Z. (2021). Lightning surge
analysis of transmission line towers with a
hybrid FDTD-PEEC method. IEEE
Transactions on Power Delivery, 37(2), 1275-
1284.
[11] Castro, W. S., Lopes, I. J., Missé, S. L., &
Vasconcelos, J. A. (2022). Optimal placement
of surge arresters for transmission lines
lightning performance improvement. Electric
Power Systems Research, 202, 107583.
[12] He, S., Xie, M., Tontiwachwuthikul, P., Chan,
C., & Li, J. (2022). Self-adapting anti-surge
intelligence control and numerical simulation
of centrifugal compressors based on RBF
neural network. Energy Reports, 8, 2434-
2447.
[13] Tian, H., Liu, H., Ma, H., Zhang, P., Qin, X.,
& Ma, C. (2021). Steady-state voltage-control
method considering large-scale wind-power
transmission using half-wavelength
transmission lines. Global Energy
Interconnection, 4(3), 239-250.
[14] Shi, J., & Oren, S. S. (2020). Flexible line
ratings in stochastic unit commitment for
power systems with large-scale renewable
generation. Energy Systems, 11, 1-19.
[15] Gusev, O. Y., Gusev, Y. P., & Posokhov, N.
O. (2023). Specific Features of Transient
Recovery Voltages during Short-Circuit
Clearing in High-Voltage Transmission Lines.
Russian Electrical Engineering, 94(1), 46-50.
[16] Tahir, M.Z., Jamaludin, R., Nawi,
M.N.M., Baluch, N.H., Mohtar, S. (2017).
Building energy index (BEI): A study of
government office building in Malaysian
public university. Journal of Engineering
Science and Technology, 12 (Special Issue
2), pp. 192-201.
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2023.18.37
M. Mohammed Thaha, Rosini Nawang Mustapen,
Rafikullah Deraman, Shanmugam Durairaj,
Rajendrakumar Ramadass
E-ISSN: 2224-350X
371
Volume 18, 2023
[17] Tahir, M.Z., Nawi, M.N.M., Rajemi, M.F.
(2015). Building energy index: A case study
of three government office buildings in
Malaysia. Advanced Science
Letters, 21 (6), pp. 1798-1801.
[18] Giraudet, F. (2020). Various benefits for line
surge arrester application and advantages of
externally gapped line arresters. Power
Research-A Journal of CPRI, 136-144.
[19] Stanchev, D. (2020, September). Energy
stress of externally gapped line arresters for
various cases through model study. In 2020
12th Electrical Engineering Faculty
Conference (BulEF) (pp. 1-4). IEEE.
[20] Agrawal, S., Sarkar, S., Aouedi, O., Yenduri,
G., Piamrat, K., Alazab, M., ... & Gadekallu,
T. R. (2022). Federated learning for intrusion
detection system: Concepts, challenges and
future directions. Computer Communications.
[21] Yildirim, E., (2020). Electricity Distribution
System Dataset. Kaggle, [Online].
https://www.kaggle.com/datasets/ensariyildiri
m/electricity-distribution-system-dataset/code
(Accessed Date: December 9, 2023).
Contribution of Individual Authors to the
Creation of a Scientific Article (Ghostwriting
Policy)
The authors equally contributed in the present
research, at all stages from the formulation of the
problem to the final findings and solution.
Sources of Funding for Research Presented in a
Scientific Article or Scientific Article Itself
No funding was received for conducting this study.
Conflict of Interest
The authors have no conflicts of interest to declare.
Creative Commons Attribution License 4.0
(Attribution 4.0 International, CC BY 4.0)
This article is published under the terms of the
Creative Commons Attribution License 4.0
https://creativecommons.org/licenses/by/4.0/deed.en
_US
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2023.18.37
M. Mohammed Thaha, Rosini Nawang Mustapen,
Rafikullah Deraman, Shanmugam Durairaj,
Rajendrakumar Ramadass
E-ISSN: 2224-350X
372
Volume 18, 2023
