Power Grid Resilience in the Face of Nuclear Electromagnetic Pulse (NEMP) Attacks
Open Access DepositedDownloadable Content
A nuclear electromagnetic pulse (NEMP, often termed HEMP) poses a serious threat to modern bulk power systems because it can produce geographically widespread cyber and physical impacts. Unlike many conventional disturbances, a high-altitude NEMP can affect exceptionally large regions within a short time span. In addition to inducing E-fields that drive geomagnetically induced current (GIC) related responses in power transmission networks, certain NEMP components can directly interfere with electronic devices, monitoring channels, and control functions. These characteristics make NEMP resilience a cyber-physical problem that requires coordinated hazard assessment and mitigation. This dissertation develops an integrated resilience framework for cyber-physical power systems (CPPS) under NEMP threats. The work is organized around three coordinated parts
geographically resolved hazard construction and assessment, cyber-impact mitigation, and physical-impact mitigation. First, a geographically resolved EMP attack E3A E-field impact assessment framework is developed for large-scale power systems. Different from workflows that begin with a prescribed E-field input, the proposed method starts from the physical formation of the disturbed patch, explicitly constructs the corresponding magnetic disturbance, derives the induced E-field in a physically consistent manner, and incorporates conductivity-aware local correction. The corrected field quantities are then mapped to substations and transmission lines to derive induced-voltage and GIC-related indicators for asset-level assessment. Second, this dissertation develops a cyber-impact mitigation method for NEMP-induced interference in CPPS measurement acquisition. An NEMP-to-power-system simulation environment is established to generate disturbance scenarios and corresponding interfered measurement datasets. Based on these data, a convolutional autoencoder is used to suppress measurement corruption, and a graph-based learning model is used to recover structured missing data by utilizing the topological relationships of the network. Third, this dissertation develops a physical-impact mitigation framework based on multi-agent deep reinforcement learning (MADRL) to support coordinated operational response under NEMP-driven disturbances. The proposed method considers mitigation actions including reactive power support, selective generator tripping, and targeted load shedding under operational constraints. Overall, this dissertation establishes a coordinated framework for NEMP resilience in power systems by connecting geographically resolved E3A assessment, cyber-impact mitigation, and physical-impact mitigation within one unified process.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.