Construction of Distribution Network Fault Detection Model Based

on Artificial Intelligence Algorithm

Uvarajan K. P., Krishnakumar S., Sridhar K., Vignesh K., Hari Sudhan M. and Ragul B. R.

Department of Electronics and Communication Engineering, K.S.R College of Engineering, Trichengode, Namakkal, Tamil

Nadu, India

Keywords: Fault Detection, Real‑Time Monitoring, Artificial Intelligence, Machine Learning, Neural Networks,

Distribution Network, Efficiency.

Abstract: The objective of this project is to create an AI-driven Distribution Network Fault Detection Model to enhance

the reliability and efficiency of the power distribution networks. Materials and Methods: The model employs

sensor information from the distribution network, treated using methods such as missing data handling, label

encoding, and Principal Component Analysis (PCA) for feature selectivity such as accuracy, F1 score,

confusion matrix, ROC, and AUC. A Graphical User Interface (GUI) is created for easy use, with users able

to upload datasets, define parameters, and visualize output. Result: The algorithms demonstrated robust

performance, with ANN and SVM delivering the best fault prediction accuracy. Conclusion: The fault

detection model using AI enhances power distribution network reliability through precise fault detection in

real-time. The performance of the model is confirmed through various metrics, and the GUI provides

simplicity in use. The system helps to optimize grid operation, minimize downtime, and maximize overall

service reliability, with possibilities for future advancements in smart grid Keywords: Fault Detection, Real-

Time Monitoring, Artificial Intelligence, Machine Learning, Neural Networks, Distribution Network,

Efficiency integration.

1 INTRODUCTION

Electricity distribution networks constitute the

backbone of the electricity network, distributing

electricity from the transmission network to industrial

complexes, commercial structures, and residential

homes. Highly complex with extensive geographical

coverage of operations, distribution networks are

susceptible to the development of faults that result in

loss of power, inefficiency, and excessive losses.

Fault detection and real-time adjustments are highly

crucial in order to fulfill the objective that the power

distribution system ought to be running at its best and

always. Fault detection previously was slow and

labor-intensive with reactive measures and extensive

testing, and this resulted in fault repair delays and

longer downtime (Nourani, Attarha, and Chakrabarty

2002). A fault detection system with the assistance of

advanced algorithms can automatically thus

significantly enhance the fault detection process with

a minimal response time and with least disruption to

the power supply(Radhoush, Whitaker, and Nehrir

2023). Machine learning (ML) and artificial

intelligence (AI) have shown much potential in the

detection of faults in power systems(Dini and Paolini

2025). The technologies can handle large volumes of

data sensed by sensors on the distribution grid and

determine outliers that are indicative of faults(“Edge

Computing with Artificial Intelligence: A Machine

Learning Perspective” 2023). Machine learning

algorithms are capable of identifying potential faults

with perfect accuracy from past fault experience and

thus reduce the need for human intervention. Some

machine learning models, such as Support Vector

Machines (SVM), K-Nearest Neighbors (KNN), and

Artificial Neural Networks (ANN), are particularly

apt for this, each of which has different strengths in

classification, scalability, and interpretability.

2 RELATED WORKS

An Artificial Intelligence (AI) algorithm-driven

Distribution Network Fault Detection Model involves

the use of advanced machine learning and data

analytics for monitoring and health checking of

P., U. K., S., K., K., S., K., V., M., H. S. and R., R. B.

Construction of Distribution Network Fault Detection Model Based on Artiﬁcial Intelligence Algorithm.

DOI: 10.5220/0013896300004919

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies (ICRDICCT‘25 2025) - Volume 3, pages

253-260

ISBN: 978-989-758-777-1

253

power distribution networks(Chen et al., n.d.). AI

algorithms such as neural networks, support vector

machines (SVM). The decision trees are trained on

real-time data collected via sensor networks and

smart grids to detect faults, forecast failures, and

enable efficient network performance(Chen et al.,

n.d.; Liwen et al., n.d.). Through the use of supervised

and unsupervised learning.

These models are capable of performing fault

classification, root-cause diagnosis, and even

disruption forecasting(Chen et al., n.d.; Liwen et al.,

n.d.; Ruirong et al., n.d.). Predictive maintenance

through artificial intelligence enables proactive

detection of impending faults before they occur(Chen

et al., n.d.; Liwen et al., n.d.; Ruirong et al., n.d.; “[No

Title],” n.d.).minimizing downtime while increasing

grid resilience(Alazemi 2024).Anomaly detection

techniques are used to identify deviation from normal

operation in the data, indicating possible faults, and

data fusion integrates inputs from diverse sources to

offer improved detection. Also, real-time monitoring

and fault finding algorithms make real-time response

possible, isolating the faulty areas of the grid for

faster restoration(Chen et al., n.d.; Liwen et al., n.d.;

Ruirong et al., n.d.; “[No Title],” n.d.; Li

2020).Therefore, fault detection using AI enhances

the efficiency and reliability of the distribution

system with minimal interruption and maintaining a

steady power supply.

3 TECHNOLOGY AND

METHODOLOGY

Simulation Software: SPICE: Its time-domain

simulation replicates the electrical behavior of the

circuit involving impedance mismatches and signal

reflection(Ru et al., n.d.). Ansys HFSS/CST Studio

products include electromagnetic field simulation

toolkits to simulate (Habib et al., n.d.)the signal

propagation and transmitting line effects such as

crosstalk and reflection. Keysight ADS finds

application in high-speed design for jitter analysis,

eye diagram, and total signal distortion. HyperLynx:

Signal and power integrity simulation at the PCB

level to assist in via and interconnect analysis.

Ansys Thermal: Thermal analysis for temperature

gradient estimation and impact on reliability and

circuit performance(Chen et al., n.d.). Methodology

Circuit Design and Setup: High-speed circuit

geometries were created with precise PCB

designing tools, i.e., microstrip traces and vias to

simulate a realistic high-frequency environment. Test

structures with different trace length(Srivastava et

al. 2022). width, and impedance were used to study

different mechanisms of signal integrity degradation.

Reliability Model: Electromigration Models:

Simulate the effects of high current density on

interconnects in order to make long-term predictions

for degradation. Aging Effects Simulation:

Simulation of degradation of material properties with

time under electrical and thermal stress to study the

effect of aging on circuit reliability(Srivastava et al.

2022; n.d.). Simulation of Signal Integrity Impedance

Matching and Reflection: Reflection due to

impedance mismatch was simulated through time-

domain simulation. Crosstalk Analysis: In order to

confirm the effects of interference, signal lines

running one beside another were simulated(Asman et

al. 2021).The impact on the signal was observed by

using an eye diagram.Jitter and Timing Analysis: For

observing the impact of timing defects on the overall

high-speed circuit behavior, jitter of a signal was

analyzed.

Group 1: The current system is founded on fault

modeling and test methods for SI analysis in high-

speed System-on-Chips (SoCs).

Group 2: The system proposed here improves SI

analysis by integrating predictive simulation

methods for degradation and reliability analysis.

4 STATISTICAL ANALYSES

Statistical methods will be utilized to contrast the

performance of the proposed AI-based fault detection

model with the conventional fault detection approach.

(Srivastava et al. 2022)Measures of performance

utilized are Accuracy, Precision, Recall, F1-Score,

Detection Speed, and False Positive Rate (FPR).A

real-time and simulated fault case data set of a

distribution network is divided into a train (70%) and

a test (30%) set.(Hossain, Rahman, and Ramasamy

2024) AI model performance is justified through k-

fold cross-validation (usually k=10) to be consistent

and not overfit. Hypothesis testing is subsequently

conducted with the Independent Samples t-test to find

a difference in the measures of the performance of the

AI model and the standard procedures. The α value

for significance level is 0.05 and the confidence to be

95%. The statistical test p-value is employed to

ensure that the AI model is producing statistically

significantly better performance(Zou et al. 2023).

Receiver Operating Characteristic (ROC) curve and

Area Under Curve (AUC) are employed to check

model performance as well.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

254

5 RESULTS

MATLAB-founded Artificial Intelligence (AI)

algorithm-founded Distribution Network Fault

Detection Model suggests the formulation of a fault

detection, diagnosis, and location system for power

distribution networks that is trustworthy. The project

starts with the acquisition of data from the sensors in

the networks and from the past records and continues

with preprocessing activities like cleaning,

normalization, and feature extraction. Machine

learning models such as decision trees, SVM, and

neural networks are used to train the models with fault

detection and identification features such as ground

faults, short-circuits, and overloads. Predictive

failures are predicted by the system beforehand and

provide real-time alerts, thereby making predictive

maintenance easier. The model indicates the exact

location of faults based on fault location algorithms

with less downtime. The project analyzes the

performance of the model based on metrics such as

precision, accuracy, and recall and renders the model

very efficient and reliable. Generally, the system

enhances grid resilience with an active network

management strategy by isolating faulty areas very

quickly and generally enhancing the stability of the

network.

6 DISCUSSIONS

The development of a fault detection AI model is a

significant shift from traditional fault detection(Zhu

et al., n.d.). The model, founded on machine and deep

learning-based algorithms, is more accurate,

dependable, and faster in detection than the

traditional approach(Zhu et al., n.d.; “Fault Diagnosis

System of Urban Power Supply and Distribution

Based on BP Neural Network,” n.d.). The enhanced

capability of the model is due to the capability of the

model to learn from historical data as well as real-

time data in a bid to enhance its capability to detect

faults. Statistical testing here in terms of Independent

Samples t-test confirms that the AI model is

statistically significantly improved with a p-value less

than the chosen significance level of 0.05. ROC curve

and AUC measures also confirm the strength and

consistency of the model. But some limitations still

exist currently, such as the need for a large(Zhou et

al. 2023) dataset to achieve high precision and

sensitivity of the model to noisy data. The

adaptability of the model needs to be enhanced in the

future and advanced AI techniques such as hybrid

models need to be employed. The real-time fault

diagnosis needs to be enhanced. The dataset needs to

be large and tested under various network conditions

to improve the model performance and

generalization.

7 CONCLUSIONS

In essence, the structure of an artificial intelligence

algorithm-based smart fault detection system in

power distribution systems is an important

enhancement of the system's reliability, effectiveness,

and sensitivity(Liang et al., n.d.). Through

incorporating the use of machine learning algorithms

such as SVM, KNN, ANN, and reinforcement

learning, the system can forecast and detect faults in

real-time, thereby enabling the preventive measure to

minimize downtime and service interruption(Liu et

al., n.d.). The application of the IoT sensors and

real-time data analysis improves the capability of the

system to make real-time prediction in a way that

faults can be detected before any extensive damage`

occurs. Flexibility and scalability are also enhanced

based on the capability of the model to compare and

contrast various algorithms(Tian et al. 2021). hence

capable of operating with various types of distribution

networks. The fault detection model itself comes to be

the foundation for smart grid technology innovations

in the future with real-time monitoring and

optimisation required. Future model developments

offer enormous opportunities for other artificial

intelligence techniques and other sources of real-time

information to be incorporated, further building its

predictive strength and fault categorization(Bai et al.,

n.d.). With the advancement of machine learning

algorithms with higher performance continuously

being developed and good sensor data to rely on, the

system can further be optimized to be more efficient,

minimize operating costs, and make edge networks

more reliable.

8 TABLES AND FIGURES

Table 1 shows the Fault Type Analysis Based on

Voltage Drop, Current Increase, Detection Accuracy,

and Fault Probability in Power Distribution

Networks. Table 2 illustrates the Statistical Summary

of Detection Accuracy in AI-Based Fault Detection

System. Table 3 represents the Results of One-

Sample T-Test on Detection Accuracy of AI-Based

Fault Detection Model.

Construction of Distribution Network Fault Detection Model Based on Artiﬁcial Intelligence Algorithm

255

Table 1: Fault Type Analysis Based on Voltage Drop,

Current Increase, Detection Accuracy, and Fault

Probability in Power Distribution Networks.

Fault

Type

Voltage

Drop(v)

Current

Increase(

Detection

Accuracy

(%)

Fault

probabi

lity

Short

Circuit

180 50 98.5 0.90

Open

Circuit

220 0 95.2 0.80

Ground

Circuit

200 30 97.0 0.85

Line-to-

line

190 40 96.8 0.87

Normal 230 5 99.9 0.05

Table 2: Statistical Summary of Detection Accuracy in AI-

Based Fault Detection System.

Detection

Accuracy

Std.

Deviati

Std.Error

Mean

81.

39.828

16.25978

Table 3: Results of One-Sample T-Test on Detection

Accuracy of AI-Based Fault Detection Model.

8.1 Flow Diagram

Figure 1: Data Preprocessing and Partitioning Workflow

for AI-Based Fault Detection Model.

External Input → Data Acquisition & Preprocessing

→ Feature Extraction → AI-Based Fault Detection

Model → sification & Localization → Decision

Support & Alerting →

External Output

Figure 2: Detailed Workflow of AI Model Training and

Fault Detection – Level 3.

Fault Detection Model of a Distribution Network

using Artificial Intelligence Algorithm development

is a process starting with data collection and

preprocessing where power grid sensor data and real-

time SCADA system data are collected, cleaned, and

preprocessed. The data is processed using feature

extracti where the significant fault indicators such as

voltage dips, frequency fluctuation, and harmonic

distortions are extracted. These features are utilized as

input to the AI fault detection-based model on

machine learning- or deep learning- for fault

detection and pattern examination. Subsequent to

fault detection, the system proceeds with fault

classification and localization for determination of

the type and location of the fault within the

distribution network. The final phase, decision

support and alerting generate comprehensive fault

reports and real-time instant alerts to the grid

operators for corrective action in a timely manner.

The model employs historic fault data and an

updatable periodic fault detection model’s database to

improve accuracy with age. the Utilizing AI, this

model significantly improves fault detection speed

and accuracy and introduces stability and credibility

into the power distribution network. Figure 1shows

the Data Preprocessing and Partitioning Workflow

for AI-Based Fault Detection Model. Figure 2 shows

the Detailed Workflow of AI Model Training and

Fault Detection – Level 3.

Detection Accuracy

Sig.(

taile

)

Differe

nce

95% Confidence

Interval

the Difference

5 .004

81.233

Lower

39.4362

per

123.03

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

256

Figure 3: AI-Based Fault Detection Flowchart (DF - Level

3) Using Machine Learning Algorithms.

Figure 3 Building a Distribution Network Fault

Detection Model Based on an Artificial Intelligence

Algorithm is a long multi-step process, with each step

carefully calibrated to ensure accuracy and reliability.

The process begins with data gathering and

preprocessing, where real-time data collected from

power grid sensors and SCADA systems are

collected, noise filtered out, normalized, and

formatted for analysis. The second one, extraction

feature, is to recognize critical fault indicators such

as voltage dipoles, harmonics, and transient

disturbances using advanced signal processing

techniques like FFT and wavelet transforms. The data

obtained after processing is then used to provide it to

the AI-based model to detect faults, where the AI

model learns from previous faults' data to extract

faults' patterns to make predictions in real-time.

9 SYSTEM ARCHITECTURE

Figure 4: Comprehensive AI-Based Fault Detection and

Feedback Workflow for Distribution Networks.

The architecture of a Distribution Network Fault

Detection Model using AI has several layers and

incorporates aspects of data acquisition, processing,

analysis, and decision-making together. The

following is a detailed explanation of the architecture

in Figure 4.

9.1 Usecase - Diagram

A Use Case Diagram facilitates the visualization of

interactions among various system components and users in

the AI-based fault detection model for a power distribution

network in Figure 5.

Figure 5: User-Initiated Workflow for AI-Based Fault

Detection in Distribution Networks.

9.2 ER Architecture

Figure 6: AI-Based Fault Detection Framework Integrating

PCA and 1D-CNN Algorithms.

ER (Entity-Relationship) Architecture establishes

how various entities in the AI-based Distribution

Network Fault Detection Model communicate with

one another. ER Architecture is a representation of

data structure and relations among various elements

of the system figure 6.

9.3 Sequence Diagram

It depicts a machine learning-based fault detection

pipeline. It starts from an erroneous data set, which is

preprocessed by missing value management and label

encoding in order to reach the data integrity level. The

data is divided between a test data set and a training

data set. Feature extraction methods such as

Component Principal Analysis (PCA) and Machine

Learning (ML) algorithms are employed to enhance

Construction of Distribution Network Fault Detection Model Based on Artiﬁcial Intelligence Algorithm

257

the performance of the model. Some such as 1D

Convolutional Neural Networks (CNNs) are trained.

After the model is trained, the model predicts based

on test data. The prediction of the model is ultimately

utilized to identify faults, which identify faults by

identifying learned patterns. Figure 7 shows the

Sequence Diagram for Machine Learning-Based

Fault Detection Process in Power Distribution

Systems.

Figure 7: Sequence Diagram for Machine Learning-Based

Fault Detection Process in Power Distribution Systems.

9.4 Activity Diagram

It can depict a fault detection model using machine

learning. It begins with a faulty dataset, which is fed

to two important steps: data preprocessing and data

partitioning. Preprocessing handles the normalization

and NCP (possibly Nonlinear Component

Processing) for cleaning and organizing the data. The

is the dataset separated into training and test sets for

the building model. The machine learning (ML)

algorithm is applied in model training to learn

patterns in the data. The trained model will then be

applied to run over the test data, which will provide

fault detection as a prediction. Fault identification is

enabled by the systematic approach and enhanced

model performance. Figure 8 shows the Workflow of

Fault Detection System Using Data Preprocessing

and ML Algorithms.

Figure 8: Workflow of Fault Detection System Using Data

Preprocessing and ML Algorithms.

10 VALIDATION PERFORMANCE

GRAPHS

10.1 Error Histogram with 20 Bins

Figure 9: Error Histogram with 20 Bins for Model

Performance Evaluation.

The 20-bin error histogram shows the distribution of

errors in a machine learning model. The x-axis is

errors, computed as targets minus outputs, and the y-

axis is the number of samples per error range. The bars

are colored with blue for training data, green for

validation data, and red for test data. There is a vertical

orange line at zero error, the location of best results.

Errors are mostly concentrated around zero, showing

good performance by the model with minor deviations

from predictions to actualities. The relative

constriction in the range of errors also serves to

highlight the precision of the model. Validation and

test errors both work towards diagnosing the capability

of the model to generalize for different sets. Figure 9

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

258

shows the Error Histogram with 20 Bins for Model

Performance Evaluation.

10.2 Output

Figure 10: Training Progress of Neural Network Model

Over 14 Epochs.

Following is training performance measurements

over 14 epochs for a machine learning model. The

first graph is the gradient decreasing consistently,

which indicates good model convergence. The second

graph is Mu, the learning rate adjustment factor,

which decreases exponentially, which indicates

optimization stable. The third graph is validation

checks, which are zero during training, which

confirms no overfitting issues. The Mu and reducing

gradient all ensure that the optimization algorithm is

successfully minimizing the loss function. The

absence of validation failures also ensures that the

model possesses a high generaliz ability to new data.

All these results indicate a flat, stable training process

without any disturbance or overfitting, ending up in

an effective trained model. Figure 10 shows the

Training Progress of Neural Network Model Over 14

Epochs.

REFERENCES

Alazemi, Talal. 2024. “Enhancing Wind Energy

Forecasting Efficiency through Dense and Dropout

Networks (DDN): Leveraging Grid Search

Optimization.” Brunel University London.

http://bura.brunel.ac.uk/handle/2438/29670.

Asman, Saidatul Habsah, Nur Fadilah Ab Aziz, Ungku

Anisa UngAmirulddin, and Mohd Zainal Abidin Ab

Kadir. 2021. “Transient Fault Detection and Location

in Power Distribution Network: A Review of Current

Practices and Challenges in Malaysia.” Energies 14

(11): 2988.

Bai, Yuling, Yunhua Li, Yongmei Liu, and Zhao Ma. n.d.

“Short-Term Prediction of Distribution Network Faults

Based on Support Vector Machine.” Accessed March

22,2025.

https://ieeexplore.ieee.org/abstract/document/8283062.

Chen, Shaomin, Yu Tian, Zhenggan Dai, Jianmin Lin,

Rumei Huang, and Haoyu Wang. n.d. “Construction of

Distribution Network Fault Detection Model Based on

Artificial Intelligence Algorithm.” Accessed March 22,

2025. https://ieeexplore.ieee.org/abstract/document/10

414966.

Dini, Pierpaolo, and Davide Paolini. 2025. “Exploiting

Artificial Neural Networks for the State of Charge

Estimation in EV/HV Battery Systems: A Review.”

Batteries 11 (3): 107.

“Edge Computing with Artificial Intelligence: A Machine

Learning Perspective.” 2023. ACM Computing

Surveys, January. https://doi.org/10.1145/3555802.

“Fault Diagnosis System of Urban Power Supply and

Distribution Based on BP Neural Network.” n.d.

Accessed March 22,2025. https://doi.org/10.1145/358

2935.3583005.

Habib, Habib Ur Rahman, Asad Waqar, Bashar Sakeen

Farhan, Tanveer Ahmad, Mehdi Jahangiri, Moustafa

Magdi Ismail, Parvez Ahmad, Asad Abbas, and Yun-

Su Kim. n.d. “Analysis of Optimal Integration of EVs

and DGs Into CIGRE’s MV Benchmark Model.”

Accessed March 22, 2025.

https://ieeexplore.ieee.org/abstract/document/9877797.

Hossain, Md Naeem, Md Mustafizur Rahman, and

Devarajan Ramasamy. 2024. “Artificial Intelligence-

Driven Vehicle Fault Diagnosis to

Revolutionize Automotive Maintenance: A Review. |

EBSCOhost” 141 (2): 951.

Liang, Jiefeng, Tianjun Jing, Huanna Niu, and Jiangbo

Wang. n.d. “Two-Terminal Fault Location Method of

Distribution Network Based on Adaptive Convolution

Neural Network.” Accessed March 22, 2025.

https://ieeexplore.ieee.org/abstract/document/9035471.

Li, H. R. 2020. “Research on Fault Location of Power

Distribution Network Based on Fault Data

Information”.

Telecommunications and Radio Engineering 79 (8).

https://doi.org/10.1615/TelecomRadEng.v79.i8.80.

Liu, Shourui, Hong Yin, Yuan Zhang, Xuan Liu, and

Chunbo Li. n.d. “Fault Location Method for

Distribution Network with Distributed Generation

Based on Deep Learning.” Accessed March 22, 2025.

https://ieeexplore.ieee.org/abstract/document/1008221

Liwen, Qin, Yu Xiaoyong, Feng Bin, Qin Zongtao, Wu

Lifang, and Ou Shifeng. n.d. “Fault Diagnosis of

Distribution Network Based on Artificial Intelligence.”

Accessed March 22, 2025.https://ieeexplore.ieee.org/a

bstract/document/9724108.n.d. Accessed March 22,

2025.https://www.iaeng.org/IJCS/issues_v47/issue_3/I

JCS 47_3_26.pdf.

Nourani, Mehrdad, Amir Attarha, and Krishnendu

Chakrabarty. 2002. “Signal Integrity: Fault Modeling

and Testing in High-Speed SoCs.” SOC (System-on-a-

Chip) Testing for Plug and Play Test Automation, 175–

90.

Construction of Distribution Network Fault Detection Model Based on Artiﬁcial Intelligence Algorithm

259

Radhoush, Sepideh, Bradley M. Whitaker, and Hashem

Nehrir. 2023. “An Overview of Supervised Machine

Learning Approaches for Applications in Active

Distribution Networks.” Energies 16 (16): 5972.

Ruirong, Ding, Liu Yinliang, Zhang Huanqiang, Yu Lei,

Xie Yi, Liang Xiaoqiang, Ge Yang, and Lin Xinhao.

n.d. “Research on Fault Location Technology of

Intelligent Distribution Network Based on Neural

Network.” Accessed March 22, 2025.

https://ieeexplore.ieee.org/abstract/document/9873778.

Ru, Jiaxin, Guomin Luo, Boyang Shang, Simin Luo,

Wenlin Liu, and Shaoliang Wang. n.d. “Fault Line

Selection and Location of Distribution Network Based

on Improved Random Forest Method.” Accessed

March 22, 2025.

https://ieeexplore.ieee.org/abstract/document/1008201

Srivastava, Ishan, Sunil Bhat, B. V. Surya Vardhan, and

Neeraj Dhanraj Bokde. 2022. “Fault Detection,

Isolation and Service Restoration in Modern Power

Distribution Systems: A Review.” Energies 15 (19):

7264.

Tian, Jingjing, Fang Geng, Feng Zhao, Fengyang Gao, and

Xinqiang Niu. 2021. “Application of Convolutional

Neural Network in Fault Line Selection of Distribution

Network.” Journal of Applied Science and Engineering

25 (1): 195–205.

Zhou, Chuan, Suying Gui, Yan Liu, Junpeng Ma, and Hao

Wang. 2023. “Fault Location of Distribution Network

Based on Back Propagation Neural Network

Optimization Algorithm.” Processes 11 (7): 1947.

Zhu, Xiaohui, Xin Lu, Dong Liu, and Bingda Zhang. n.d.

“An Improved Fault Locating System of Distribution

Network Based on Fuzzy Identification.” Accessed

March 22,2025. https://ieeexplore.ieee.org/abstract/do

cument/5735873.

Zou, Xiuguo, Wenchao Liu, Zhiqiang Huo, Sunyuan Wang,

Zhilong Chen, Chengrui Xin, Yungang Bai, et al. 2023.

“Current Status and Prospects of Research on Sensor

Fault Diagnosis of Agricultural Internet of Things.”

Sensors 23 (5): 2528.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

260