Point estimate method based on univariate dimension reduction model for probabilistic power flow computation

This study employs Gaussian copula to model correlated non‐normal variables in power system, whereby the probabilistic power flow (PPF) problem is transformed to independent standard normal space. In conjunction with univariate dimension reduction model, two quadrature rules: Gauss‐logistic (GL) quadrature and Clenshaw–Curtis (CC) quadrature, are developed to calculate the moments of PPF solutions; for CC quadrature, the weights and nodes are given explicitly. Testing on a modified 118‐bus system, it is found that CC quadrature converges more uniformly than the generally used Gauss–Hermite quadrature, and GL quadrature is more accurate.