Statistic linear parametric techniques for residential electric energy demand forecasting. A review and an implementation to Chile

In operational and planning studies in power electric distribution systems, one of the most important tasks is to quantify the evolution of the system. In particular, it is necessary to be able to measure the growth of electrical demand, with special attention to residential consumption. For that reason, it is fundamental to predict its future values. Considering the availability of real measure data, statistic parametric methods are widely used to describe and forecast those residential loads. This paper reviews the principal statistical linear parametric methods and implements four of them to analyse real measure data from Chilean systems. Additionally, those methods are compared among them and the performance of a non-tested continuous approach based on diffusion processes can be evaluated. In each case, the parametric adjustment and the validation methods are explained.