Data mining techniques applied to spatial load forecasting
Franz Henry Pereyra Zamora, Carlos Márcio Vieira Tahan · 2005
Load forecasting in distribution systems is essentially important and it is a challenge due to its spatial diversity associated to the land-use and consumers that inhabit on it. Many factors should be considered, such as: which type of land-use exists and will exist and which type of energy consumption the territory in analysis has or will have; when new substations and new feeders should be built or when existent facilities should be reinforced; where to plan new lines and structures. This work presents the data mining techniques applied to a methodology for spatial load forecasting which is aimed to the planning of distribution systems. This methodology addresses aspects such as: consumers' data, network load data, end-use load curve standards, consumers load curves aggregation, small areas, classification of electric loads and patterns recognition. (4 pages)