Scholarlink

<p>This paper analyzes the impact of load redistribution (LR) attacks on power distribution systems regarding significant operational constraints and vulnerabilities. LR attacks, a form of False Data Injection Attacks (FDIAs), manipulate load measurements, disturbing state estimation and causing severe operational disruptions. In response to this threat, we propose a resilient detection and mitigation scheme based on SVMs. The suggested method leverages real-time data analysis and advanced machine learning algorithms to enhance cyberattack resilience. Large-scale systematic simulations on the IEEE 123-node test case validate the performance of the proposed methodology, sustaining a 98.5% detection rate while significantly mitigating the adverse effects of LR attacks on voltage stability and power flow efficiency. The research identifies the potential of machine learning (ML)-based systems to enhance power distribution system security against cyber threats.</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

Detection and Mitigation of Load Redistribution Attacks in Power Distribution Systems Using Machine Learning

  • 期刊论文

  • 发表日期：

    2026年1月28日

查看全文
  • Youjun Qi,
  • Tianren Cha,
  • Suyan Zhang

摘要

This paper analyzes the impact of load redistribution (LR) attacks on power distribution systems regarding significant operational constraints and vulnerabilities. LR attacks, a form of False Data Injection Attacks (FDIAs), manipulate load measurements, disturbing state estimation and causing severe operational disruptions. In response to this threat, we propose a resilient detection and mitigation scheme based on SVMs. The suggested method leverages real-time data analysis and advanced machine learning algorithms to enhance cyberattack resilience. Large-scale systematic simulations on the IEEE 123-node test case validate the performance of the proposed methodology, sustaining a 98.5% detection rate while significantly mitigating the adverse effects of LR attacks on voltage stability and power flow efficiency. The research identifies the potential of machine learning (ML)-based systems to enhance power distribution system security against cyber threats.