Load profile estimation in electric transmission networks using independent component analysis

Abstract

It is important to estimate electric loads profiles in the deregulated environment where competing entities need to assess the load demands based on partial knowledge of the system. Independent component analysis (ICA) is a statistical technique used to separate linear mixtures of statistical independent source signals by maximization of negentropy. In this paper, we apply ICA to estimate load profiles using only a small set of active line flow measurements without prior knowledge of the electric network model parameters or topology. A filtering based preprocessing technique is used to ensure statistical independence of load components. The influence of measurement noise and nonlinearity of the power flow model are also investigated. The proposed approach is demonstrated for a five-bus system as well as the IEEE 30-bus system.