High-Performance Computing-Driven Defense Architecture Against Cross-Domain Cyber- Physical Attacks in Smart Power Grids

Power infrastructure has evolved into an intelligent cyber-physical system (CPS) characterized by high levels of informatization, automation, and computationintensive operations. While this integration enhances the efficiency and real-time performance of power systems, it also introduces new cybersecurity challenges, particularly cross-domain attacks that originate in the information domain and propagate into the physical domain. Such attacks can lead to systemic paralysis and large-scale blackouts. Traditional defense mechanisms that operate in isolation on either the cyber or physical side lack the scalability and responsiveness required to address these threats. This paper proposes a high-performance computing (HPC)- enabled cyber-physical collaborative defense framework, augmented by artificial intelligence, to mitigate cross-domain attacks. Leveraging HPC resources allows for accelerated data processing, real-time anomaly detection, and dynamic system modeling across large-scale, distributed power infrastructure. The framework develops defense strategies across three dimensions: perception, identification, and disruption, each aligned with different stages of attack propagation. Case studies using simulation environments demonstrate that the proposed AI- and HPC-driven defense architecture can effectively enhance the resilience and operational stability of modern power systems.

Hali tarjima qilinmagan