Statistical and probabilistic models for smart electricity distribution networks

Abstract

This thesis aims towards an improved statistical understanding of distribution feeders, an improved probabilistic understanding of loads and methods for network-wide assessment of SmartGrid technologies. An efficient multi variable statistical analysis method was presented to identify prototypical feeders, which relies upon a few key variables that are highly meaningful from an engineering perspective and readily available in most distribution companies. Hybrid models for residential consumer load were built for high and low demand days.