A Quantitative Risk Assessment Method for Power Grids Based on High-Dimensional Information from Random Matrix Theory in Extreme Scenarios

Traditional power grid analysis methods, which rely on simplified assumptions and physical mechanism models, are effective in describing system behavior under normal operating conditions. However, they fail to fully capture the system's nonlinear coupling effects and random variations under extreme scenarios, making it difficult to cope with the complexity and uncertainty brought about by the increasing share of distributed resources. To address these issues, this paper proposes a power grid risk assessment method for extreme scenarios based on high-dimensional statistical analysis and random matrix theory. The method combines high-dimensional spatiotemporal data modeling, generative adversarial network - based data generation, and high-dimensional random matrix LES eigenvalue indicators to break through the limitations of traditional grid analysis methods. It is able to provide more accurate risk assessment under complex conditions such as extreme weather and sudden accidents.