Clasificación de los usuarios del servicio eléctrico a partir de los datos de telemedición comercial e industrial en el área de concesión de la Empresa Eléctrica Ambato Regional Centro Norte S.A (EEASA).

Abstract

This titling project consists of obtaining demand curves to be introduced into the ADMS, with the aim of improving the state estimator, for industrial and commercial users within the concession area of the EEASA (Empresa Eléctrica Ambato), to for this, classification algorithms with machine language are used for the treatment of each of the data. The methodology consists of four stages. The first stage begins by downloading each consumption profile with telemetering data from the 184 commercial and industrial clients from the EEASA portal, later this data is stored in an Excel sheet. The second stage is based on the ordering, cleaning and classification of the historical data of commercial and industrial type of electricity consumption, to then start with the pre-processing by applying the K-means algorithm, which is an unsupervised classification algorithm (clustering), thus forming a data matrix that is later exported to Python. The third part consists of the use of PCA's (Principal Component Analysis) for the reduction of the large volume of data, immediately these data are characterized. Finally, the application of the Fuzzy-c-means algorithm to obtain a quarterly atypical curve, obtaining as a result the demand curves that will serve as a reference for the behavior of industrial and commercial users.