抄録
Against the backdrop of high-penetration renewable energy integration and the convergence of wide-area sensing and intelligent control, power systems are rapidly evolving into highly coupled cyber–physical systems, whose security challenges have expanded from traditional physical instability to systemic risks driven by cross-domain cyber–physical interactions. Compared with isolated information security or equipment reliability problems, cyber–physical power system threats exhibit multidimensional attack surfaces, cross-layer propagation, covert manipulation, and coordinated cyber–physical amplification, posing fundamental challenges to state awareness, control consistency, and system stability. This paper systematically reviews recent advances in cyber–physical power system security by summarizing the architectures, attack pathway evolution, and modeling paradigms of representative threats such as false data injection, industrial communication attacks, and coordinated cyber–physical attacks, and by comparatively analyzing attack detection, situational awareness, and risk assessment methods based on physical models, data-driven approaches, and cyber–physical fusion frameworks, together with their applicability and limitations in interpretability, real-time performance, and robustness. From a system-level perspective, it further examines progress in active defense, resilient control, and security-oriented resilience enhancement, highlights the shift from passive protection to adaptive and evolutionary resilience, and finally establishes a unified analytical framework centered on “threat modeling–key technologies–system resilience evolution,” while outlining future directions including AI-enabled autonomous defense, digital twin–driven security assessment, and resilience-oriented system design to support both theoretical research and engineering practice.
