Obtención de curvas horarias de carga características desagregadas por tipo de consumidor mediante la técnica de agrupamiento espacial basado en densidad.

Abstract

The obtaining of characteristic curves by type of consumer is approached using a grouping algorithm in Matlab. The curves are obtained using measurements of the quality teams of the Azogues Electric Company (EEA). The information is prepared in MS Excel and an ordered database with a homogeneous format is generated for the analysis. In addition, the Matlab numerical computing system is used, which imports the database and through an iterative grouping algorithm based on density, generates daily demand curves for each measurement, with a breakdown of weekdays and weekends, identifying them by type of consumer. The input arguments of the Matlab Density Based Spatial Clustering Algorithm with Noise (DBSCAN) function require Euclidean distances, as well as their shape to group them considering the similarity of the studied curves. Finally, the groups obtained are validated by means of a Silhouette verification index (IS) and bad groupings are identified to later determine what the results are, which are presented by means of tables with hourly values on weekdays and weekends. The graphical comparison of the resulting curves with the previous methodology of real measurements is also added.