Generating Synthetic Power Grids using Exponential Random Graph Models

A major problem arising when dealing with power systems is a substantial lack of data due to their confidential nature. The goal of this thesis is to overcome this issue by generating new synthetic power networks using a specific class of random graph models (ERGM). In order to do this we developed both the theory to justify our assumption and fit the models as well as an efficient computational procedure to estimate them. In this work we used tools and techniques from Stochastic Processes theory, High Dimensional Probability, Graph Theory, High Dimensional Statistics, Network Analysis from a statistical and a topological point of view.