A probabilistic framework for assessing and enhancing the resilience of power systems to typhoon hazards

Typhoon is a major cause of power grid disruptions, leading to widespread blackouts and substantial financial losses. The uncertainty of typhoon events challenges grid reinforcement efforts, but most research focuses on a single typhoon scenario, neglecting the uncertainty in regional typhoon events. Additionally, few studies propose a comprehensive framework that covers the full process from regional typhoon risk modeling to power grid resilience assessment and enhancement. This paper addresses these gaps by presenting a complete framework for power system resilience assessment and enhancement under regional typhoon hazards. The framework integrates an infrastructure-centric full track model (ICFM), a hazard quantization (HQ) optimal sampling method, an artificial neural network (ANN) surrogate model and a Bayesian optimization (BO) approach. Results demonstrate that ICFM improves typhoon intensity simulation accuracy compared to conventional methods within the target area. The HQ and ANN methods enable effective resilience assessment, while BO shows superior efficiency to traditional genetic algorithms for resilience enhancement. The results confirm that this framework is an effective tool for assessing the resilience of power systems, and it lays the foundation for efficiently formulating resilience enhancement strategies for transmission towers within the power system under resource constraints.