Probabilistic steady-state analysis of power systems

Abstract

This project proposes a probabilistic load flow approach based on Gaussian process regression. The objective is to evaluate the power flow solutions in the presence of wind farms with uncertain wind energy outputs. The technique rely on input information such as prior and point values to compute posterior predictive distribution. The results are compared with that of Monte-Carlo simulation in terms of the accuracy and the computational time. Numerical simulations with IEEE 30-bus system demonstrate the feasibility of the proposed method.