A Method to Analyzing and Clustering Aggregate Customer Load Profiles Based on PCA

The determination of secondary substations load profiles may facilitate distribution system operators to better forecast, plan and operate their distribution networks. The changing in load and coincidence factors due to the high penetration of distributed generators and loads such as electric vehicles, heating ventilation and air conditioning systems, induction stoves, highlights that standard load profiles are not adapted to the current evolution of the power system. As a result, load profiles analysis has become more significant and valuable. This paper deals with the analysis and clustering of aggregate load profiles by means of principal component analysis. Starting from an extensive field measurement-based database of secondary substations daily load profiles, we used principal component analysis to extract the main components and to reduce data dimension. Thanks to the most significant principal components secondary substations are groups in homogeneous clusters labelled with a standard load profile. The proposed methodology was applied to real load profiles gathered from UNARETI, the distribution system operator of Milano.