Global Sensitivity Analysis for Rare Events in Probabilistic Power Flow Using Subset Simulation

Abstract

Traditionally, global sensitivity analysis (GSA) measures the importance of renewables and loads that will affect all probabilistic density functions (PDF) of operating states in power systems. However, when it comes to the rare events that better quantify the risk probabilities in the PDF tails, the traditional GSA is no longer applicable. In addition, the model uncertainties of the input PDFs are ignored in the traditional GSA. To overcome these challenges, we propose a novel GSA framework that addresses the probability of rare events associated with the distribution parameters of the input PDFs. Moreover, we propose to accelerate the expensive rare event-based GSA (REGSA) via the subset simulation (SS) method. The simulation results reveal the excellent performance of the proposed method.