Data analytics and machine learning-based stability assessment of active grids

Abstract

This project focuses on using Gaussian Process (GP) as a machine learning tool to solve Probabilistic Optimal Power Flow for systems with load uncertainties and renewable sources. It also tests the accuracy and competency of GP-POPF, by the use of different kernels, under the different number of bus systems. With results obtained with the use of GP for POPF, they were compared to results obtained from the traditional use of Monte-Carlo Simulations (MCS) with the purpose of minimizing error measurements