Scholarlink

This paper presents a novel integration of principal component analysis (PCA) for attack detection with adaptive sliding mode control (SMC) for mitigating cyber-physical attacks (CPAs) in power systems. Key contributions include: (1) an enhanced PCA framework with temporal correlations and adaptive thresholds achieving detection response times of  ± 0.01 s; (2) an advanced SMC incorporating dynamic boundary layers and gain adaptation mechanisms that ensures robust performance under attack conditions; (3) a computationally efficient implementation suitable for real-time deployment. Simulations on an IEEE 3-machine 9-bus system demonstrate frequency deviations contained within  ± 0.15 Hz and rotor angle stability maintained within  ± 2.25°. The integrated approach achieves 0.93 system resilience while reducing settling time and oscillation magnitude by approximately 30% compared to conventional control methods, demonstrating practical viability for enhancing power system security.

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 Using PCA – Adaptive Sliding Mode Controller

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Seema Yadav,
  • Nand Kishor,
  • Shubhi Purwar

摘要

This paper presents a novel integration of principal component analysis (PCA) for attack detection with adaptive sliding mode control (SMC) for mitigating cyber-physical attacks (CPAs) in power systems. Key contributions include: (1) an enhanced PCA framework with temporal correlations and adaptive thresholds achieving detection response times of  ± 0.01 s; (2) an advanced SMC incorporating dynamic boundary layers and gain adaptation mechanisms that ensures robust performance under attack conditions; (3) a computationally efficient implementation suitable for real-time deployment. Simulations on an IEEE 3-machine 9-bus system demonstrate frequency deviations contained within  ± 0.15 Hz and rotor angle stability maintained within  ± 2.25°. The integrated approach achieves 0.93 system resilience while reducing settling time and oscillation magnitude by approximately 30% compared to conventional control methods, demonstrating practical viability for enhancing power system security.