Paper Conference

Proceedings of eSim 2022: 12th Conference of IBPSA-Canada

     

A clustering framework to identify non-residential building archetypes using subsystem-level targeting

Rajeev Kotha, François Lédée, Mohammad Haris Shamsi, Ralph Evins
Energy in Cities group, Department of Civil Engineering, University of Victoria, Victoria, BC, Canada

Abstract: With the proliferation of building energy management systems and smart meters, high-resolution time-series data have become more easily accessible. The commercial building stock is often represented through aggregated representation that do not consider the time-series profiles. It is crucial to determine the types of patterns in these profiles to provide a deeper understanding of the building’s operations. This study proposes a clustering framework that uses time-series trends to identify commercial building archetypes. The case study uses synthetic time-series data generated using US DOE building archetype EnergyPlus models. Results indicate that water heater use side outlet temperature as a feature is effective in formation of a building cluster which contains buildings that closely resemble a medium office building archetype (homogeneity score = 0.8). Stakeholders could use the identified clusters to model deep retrofits or perhaps run targeted energy-efficiency campaigns.
Keywords: data-driven, clustering, time-series, archetypes, building management system
Paper:
esim2022_271