Customers' demand clustering analysis — A case study using smart meter data

摘要

With the addition of smart measurement devices to the electrical distribution system, it is possible to record consumption data every hour of the day, or even in fractions of an hour, for example 30 minutes, for each customer of the electric utility. This high data granularity could be used to determine the customers' consumption behavior at a daily scale. Clustering analysis of residential customers' daily demand can help allocating these profile into homogeneous groups. Then, an analysis of seasonality and variations in the behavior of each stratum is studied for the most characteristic periods of demand during the year. This study will aid to develop planning strategies for distribution systems, as in the design of stratification according to their daily consumption behavior.