Scholarlink

Smart grids, which combine powerful information technology with energy infrastructure, are revolutionizing the distribution of energy in the field of Cyber-Physical Systems (CPS). However, creative solutions are required because of their vulnerability to cyber threats. In order to provide improved cybersecurity, this work applies federated learning (FL) for malware detection in smart grids. Local models based on Supervisory Control and Data Acquisition (SCADA) use deep neural networks in a two-tier approach to identify malware context-aware. By combining information from local models, federated learning creates a global model that protects privacy. The first dataset confirms the dependability of the SafeGrid Classifier with a remarkable accuracy of 94.5% and recall of 92%. This highlights the versatility of FL with a progressive performance ascent.

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

Cyber Defense in Energy: Federated Learning Solution for Malicious Intrusions in Smart Grids

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Saira Shafi,
  • Noshina Tariq,
  • Muhammad Ashraf,
  • Ghadah Alwakid,
  • Mamoona Humayun,
  • Momina Shaheen

摘要

Smart grids, which combine powerful information technology with energy infrastructure, are revolutionizing the distribution of energy in the field of Cyber-Physical Systems (CPS). However, creative solutions are required because of their vulnerability to cyber threats. In order to provide improved cybersecurity, this work applies federated learning (FL) for malware detection in smart grids. Local models based on Supervisory Control and Data Acquisition (SCADA) use deep neural networks in a two-tier approach to identify malware context-aware. By combining information from local models, federated learning creates a global model that protects privacy. The first dataset confirms the dependability of the SafeGrid Classifier with a remarkable accuracy of 94.5% and recall of 92%. This highlights the versatility of FL with a progressive performance ascent.