Two-stage Clustering for Profiling Residential Customer Demand

Abstract

Since the power system operation and planning depend on generation and consumption behavior, load and generation patterns are fundamental inputs for power system analyses. TSOs and DSOs use typical daily profiles for representing the consumption of the end-users. The evolution of power systems, due to the increased integration of renewables, and the changed end-users practices are stressing the operation and planning processes. The updating of the typical load profiles to the behavior of current customers is often disregarded, and currently the used profiles refer to out-of-date and incomplete measurement campaigns. This paper proposes an innovative two-stage clustering methodology, able to find typical load profiles of residential customers. The focus to the residential customers is due to their extremely variable behavior in their consumption. The results add to the current practice useful improvements for planning and operation studies.