Spatiotemporal Data Mining for Distribution Load Expansion

Abstract

The load spatial forecasting is fundamental for the electric energy distribution systems planning. Several methods using different conceptions have been proposed to determine the future configuration of the electric markets. This paper proposes a dynamic model of load expansion, based on concepts of local analysis using ideas and applications from urban poles theory. Thus, the load expansion is simulated in a dynamic way, maintaining a continuous change in the conditions for localization of a new load unit. An algorithm generating a snapshot that represents the distribution system configuration at that instant determines the geometry of the market in a given instant. The proposed dynamic model, based on the urban poles theory, has the capacity for summing up the information from economic variables sets, expressed in terms of interchange flow laws, which are modeled by distance and transportation functions. This supplies the model with the capacity for being used even though the number of available explanatory variables is reduced.

Keywords

• data mining
• intelligent systems
• load modeling
• power distribution
• urban areas