In recent years, many buildings have been fitted with smart meters, from which high-frequency energy data is available. However, extracting useful information efficiently has been imposed as a problem in utilizing these data. In this study, we analyzed district heating smart meter data from 61 buildings in Copenhagen, Denmark, focused on the peak load quantification in a building cluster and a case study on load shifting. The energy consumption data were clustered into three subsets concerning seasonal variation (winter, transition season, and summer), using the agglomerative hierarchical algorithm. The representative load profile obtained from clustering analysis were categorized by their profile features on the peak. The investigation of peak load shifting potentials was then conducted by quantifying peak load concerning their load profile types, which were indicated by the absolute peak power, the peak duration, and the sharpness of the peak. A numerical model was developed for a representative building, to determine peak shaving potentials. The model was calibrated and validated using the time-series measurements of two heating seasons. The heating load profiles of the buildings were classified into five types. The buildings with the hat shape peak type were in the majority during the winter and had the highest load shifting potential in the winter and transition season. The hat shape type’s peak load accounted for 10.7% of the total heating loads in winter, and the morning peak type accounted for 12.6% of total heating loads in the transition season. The case study simulation showed that the morning peak load was reduced by about 70%, by modulating the supply water temperature setpoints based on weather compensation curves. The methods and procedures used in this study can be applied in other cases, for the data analysis of a large number of buildings and the investigation of peak loads.
A Conceptual Framework for Minimizing Peak Load Electricity using Internet of Things