The energy utilities, for ensure that your network be reliable, need to perform a procedure for study and analysis based in your functions of delivery of energy in the points of the consumption. This study, generally called of systems planning of electric power distribution, is essential for ensure that variations in the energy demand doesn’t affect the system performance, that should whether keep operating of technique manner and viable
economically. In these studies are generally analyzed, demand, typology of load curves, load factor and other aspects of the existing loads. Considering then the importance of the determining of the typologies of load curves for utilities in their planning process, the Electricity Company of Amapá (CEA) conducted a campaign of measures of load curves of the distribution transformers that were utilized for obtainment of the typologies of load curves that characterize your consumers. In this paper presents the satisfactory results obtained as from the utilization of Data Mining based in Computational Intelligence (Self-Organizing Maps of Kohonen) for selection of the typical curves and determination of the typologies of load curves of residential and industrial consumers for the city of Macapá, located in the state of Amapá. The self-organizing map of Kohonen is a type of artificial neural network that combines operations of projection and clustering, allowing the realization of exploratory data analysis, with the goal of producing summarized descriptions of large data sets.
|
