Probability Assessment of Power System Resilience Based on Stochastic Dynamics Under Sandstorms

Frequent extreme sandstorm disasters have led to power system failures and significant losses, drawing widespread international attention. Considering the impact of stochastic disturbances is crucial for assessing system resilience. This paper proposes a resilience probabilistic assessment method based on stochastic dynamics specifically for extreme sandstorm events. First, the effects of sandstorms on various aspects of the power system are examined, and a system failure model is developed. Next, a fast generative adversarial network (Fast GAN) is introduced to extract relevant features from historical sandstorm disaster data, enabling the generation of extreme disaster scenarios. Then, the Itô process from stochastic dynamics is applied to describe the system’s stochastic disturbances, and resilience metrics are constructed to obtain the probabilistic distribution of system resilience indicators. Finally, a modified IEEE 14-bus system is used in a case study to validate the feasibility and effectiveness of the proposed method in generating extreme scenarios and assessing resilience probabilities.