Scholarlink

The demand for power supply is ever-increasing, and better system performance and power quality have led to the development of the automation of technical processes. The distributed generation system and microgrid are designed to be self-sufficient in smart grid operation, and their reliability is enhanced by self-automation. This paper presents the application of dependable machine learning (ML) frameworks for the identification and categorization of series and shunt faults in distribution networks. The MATLAB simulation is used to model the IEEE 15 bus radial distribution system (RDS) and generate system data. This data is further used to train ML models by introducing various types of faults in the network. The ML techniques, primarily decision tree (DT), narrow neural network (NNN), and support vector machine (SVM), are used to identify and categorize faults in the distribution network. The efficacy and precision of multiple models were compared to determine the best fitted technique for fault identification and classification. The majority of the existing literature in this domain has demonstrated encouraging progress in identifying faults in transmission lines. However, there is a research gap when it comes to detect and classify faults on the distribution side. This paper has addressed the above-mentioned gap, and the results confirm the effectiveness of the suggested techniques.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Comparative Analysis of Fault Detection and Classification in Distribution System Using Machine Learning Techniques

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Ankur Kumar,
  • Niraj Kumar Choudhary,
  • Nitin Singh

摘要

The demand for power supply is ever-increasing, and better system performance and power quality have led to the development of the automation of technical processes. The distributed generation system and microgrid are designed to be self-sufficient in smart grid operation, and their reliability is enhanced by self-automation. This paper presents the application of dependable machine learning (ML) frameworks for the identification and categorization of series and shunt faults in distribution networks. The MATLAB simulation is used to model the IEEE 15 bus radial distribution system (RDS) and generate system data. This data is further used to train ML models by introducing various types of faults in the network. The ML techniques, primarily decision tree (DT), narrow neural network (NNN), and support vector machine (SVM), are used to identify and categorize faults in the distribution network. The efficacy and precision of multiple models were compared to determine the best fitted technique for fault identification and classification. The majority of the existing literature in this domain has demonstrated encouraging progress in identifying faults in transmission lines. However, there is a research gap when it comes to detect and classify faults on the distribution side. This paper has addressed the above-mentioned gap, and the results confirm the effectiveness of the suggested techniques.