Scholarlink

<p>Diagnosing high-impedance faults (HIFs) in electrical distribution systems (DSs) poses a significant challenge for researchers worldwide. Among the types of faults that can occur in DSs, HIFs arise from the contact between an energized conductor and a high-impedance surface. Due to their low fault current, conventional methodologies often struggle to accurately diagnose HIFs, which includes their detection, classification, and location. Most existing research focuses on fault detection, and recent research has been focused on HIF location. However, the faulty phase classification is hardly present in the literature, but it can be imperative for improving HIF diagnosis. Hence, this paper proposes an investigative analysis to determine the most suitable features and decision methods based on intelligent algorithms for HIF classification in DSs with distributed generation (DG). The study encompasses the influence of noise and the variation of DG power. The feature selection analysis demonstrates the high correlation between voltage-based features and the faulty phase. The decision tree, random forest, and k-NN algorithms achieved high accuracy even with noisy input signals and DG power variation. Overall, the study presents a new methodology to classify faulty phase when HIFs occur and provides researchers with tools to improve their diagnosis in DSs.</p>

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

High-Impedance Faults Phase Classification Using Intelligent Algorithms

  • 期刊论文

  • 发表日期：

    2025年8月2日

查看全文
  • Gabriela Nunes Lopes,
  • Maurício Pavani da Silva,
  • José Carlos de Melo Vieira,
  • Rafael Ris-Ala

摘要

Diagnosing high-impedance faults (HIFs) in electrical distribution systems (DSs) poses a significant challenge for researchers worldwide. Among the types of faults that can occur in DSs, HIFs arise from the contact between an energized conductor and a high-impedance surface. Due to their low fault current, conventional methodologies often struggle to accurately diagnose HIFs, which includes their detection, classification, and location. Most existing research focuses on fault detection, and recent research has been focused on HIF location. However, the faulty phase classification is hardly present in the literature, but it can be imperative for improving HIF diagnosis. Hence, this paper proposes an investigative analysis to determine the most suitable features and decision methods based on intelligent algorithms for HIF classification in DSs with distributed generation (DG). The study encompasses the influence of noise and the variation of DG power. The feature selection analysis demonstrates the high correlation between voltage-based features and the faulty phase. The decision tree, random forest, and k-NN algorithms achieved high accuracy even with noisy input signals and DG power variation. Overall, the study presents a new methodology to classify faulty phase when HIFs occur and provides researchers with tools to improve their diagnosis in DSs.