Key points are not available for this paper at this time.
The rapid penetration of renewable energy resources has introduced significant uncertainty into modern power systems, necessitating accurate pre-assessment to ensure secure operation. Probabilistic power flow (PPF) analysis is a powerful technique for quantifying such uncertainty, but its reliance on repeated AC power-flow solutions makes it computationally prohibitive. This study proposes a fast PPF framework that couples the linear DC power flow (DC method) model with an Extreme Learning Machine (ELM) surrogate. By augmenting ELM inputs with DC-PF results and employing Latin Hypercube Sampling, the method achieves both high speed and high accuracy. Numerical experiments on benchmark IEEE systems confirm that the proposed approach preserves the precision of conventional Monte-Carlo-based PPF while reducing total computation time by approximately 140 times.
Ohsawa et al. (Sat,) studied this question.