Scholarlink

This article introduces a novel approach to anomaly detection in cable systems, addressing the criticality of maintaining electrical system integrity. Utilizing advanced deep learning methods, specifically autoencoders, the suggested approach employs neural networks to compute reconstruction errors, enabling accurate detection of anomalies, such as the quench phenomenon. Experimental validation using real-world coaxial cable data, with simulated faults resembling actual defects, demonstrates the approach’s efficacy in enhancing stability and reliability within coaxial cable systems. Comparative analysis indicates superior results for single and multiple faults in cables, highlighting the potential of the suggested algorithm to advance cable diagnostic methods significantly and enhance overall system maintenance.

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

Anomaly Detection in Co-axial Cable Using Time–Frequency Domain Reflectometry and Deep Learning Technique

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Ashish Raj,
  • Subrata Karmakar

摘要

This article introduces a novel approach to anomaly detection in cable systems, addressing the criticality of maintaining electrical system integrity. Utilizing advanced deep learning methods, specifically autoencoders, the suggested approach employs neural networks to compute reconstruction errors, enabling accurate detection of anomalies, such as the quench phenomenon. Experimental validation using real-world coaxial cable data, with simulated faults resembling actual defects, demonstrates the approach’s efficacy in enhancing stability and reliability within coaxial cable systems. Comparative analysis indicates superior results for single and multiple faults in cables, highlighting the potential of the suggested algorithm to advance cable diagnostic methods significantly and enhance overall system maintenance.