On the Study of Commercial Losses in Brazil: A Binary Black Hole Algorithm for Theft Characterization

Abstract

According to The Brazilian Electricity Regulatory Agency, Brazil reached a loss of approximately U.S.$4 billion in commercial losses during 2011, which correspond to more than 27 000 GWh. The strengthening of the smart grid has brought a considerable amount of research that can be noticed, mainly with respect to the application of several artificial intelligence techniques in order to automatically detect commercial losses, but the problem of selecting the most representative features has not been widely discussed. In this paper, we make a parallel among the problem of commercial losses in Brazil and the task of irregular consumers characterization by means of a recent meta-heuristic optimization technique called Black Hole Algorithm. The experimental setup is conducted over two private datasets (commercial and industrial) provided by a Brazilian electric utility, and it shows the importance of selecting the most relevant features in the context of theft characterization.