Electricity demand forecasting 2030 by decomposition analysis of open data

Abstract

The demand of electrical energy in the household sectorfollowed a nearly linear growth trend for a long timemaking demand forecasting relatively simple. However, inthe last decade the growth has stalled due to energyefficiency policies, structural changes in the society andemergence of new technologies. In sparsely populatedareas the population is continually declining which affectselectrical energy consumption and increases averageconductor length per customer. These changes in theoperational environment pose challenges to demandforecasting. Historical data relating to the change factorscould be used to improve demand forecasts. This studyintroduces a method that uses decomposition and timeseriesanalysis of open data to forecast future electricalenergy demand. The method is used to forecast theelectrical energy consumption for the household sector ina group of Finnish municipalities which have a decliningpopulation.