Load modelling in commercial power systems using neural networks

Abstract

Power system load modelling is of vital importance in power flow, transient stability and voltage stability studies. It is, however, a very difficult task because load representation is qualitatively different in many aspects. Conventional approaches employ mathematical models to represent the steady and dynamic characteristics of various loads. With the advent of neural computing, attempts have constantly been made to address this problem by using this new technique. This paper discusses the applications of neural networks to the representation of the aggregation of busbar loads which are comprised of mixed but known composition.<>