Self-Organizing Maps for Classification of the Rio De Janeiro State Cities Based on Electrical Energy Consumption

The purpose of the present work is to classify the 31 cities of Rio de Janeiro State in Brazil based on their energy consumption. The point is to search new criteria to cluster the users in order to establish, in a more homogeneous way, indices of energy quality. Moreover, it aims to bring about a framework from which it will be possible to determine the relative efficiency among the cities of all Brazilian states. Traditionally this classification task is carried out using a statistical technique known as K-means, in which only five variables are considered: the size of the main network in Kilometres, the offered power, the number of users, the average monthly consumption and the covered area. This paper uses the Kohonen Self Organizing Maps technique applied to 21 variables, including the residential, industrial, public and rural consumptions in order to seek a better classification.