Characterizing Meteorology-Driven Flexibility Dynamics in Power Systems with High Renewable Integration

This study proposes a Particle Swarm Optimization (PSO)-optimized Back Propagation Neural Network (BP neural network) method for predicting power system flexibility requirements under high renewable energy integration. Departing from traditional methods that separately model generation and load, this approach establishes a direct meteorological-to- flexibility mapping framework. It enables reducing error propagation and enhancing interpretability under weather- induced uncertainty, and demonstrates significant improvements in prediction robustness and adaptability under complex meteorological conditions. Furthermore, experimental results show that this model exhibits robust performance in extreme value prediction for most temporal scales. The proposed method provides a more reliable quantitative basis for power system dispatch decisions, offering valuable insights for the construction and operation of power systems with high renewable energy penetration.