Deriving high-fidelity residential building archetypes and typical usage patterns from national energy use surveys to enhance “initial guesses” for Urban Building Energy Model (UBEM) inputs

Urban Building Energy Models (UBEMs) play a crucial role in improving energy efficiency and supporting decarbonization efforts across regions. However, scalability remains a challenge due to the lack of building-level information on construction, systems properties, and usage patterns. This study presents a novel method to derive detailed residential building archetypes from open urban datasets and the national Residential Consumption Survey (RECS), introducing a modular framework for archetype classification, assignment, and characterization. As a case study for residential buildings in climate zone 6A in the US, this study defines 41 sub-archetypes in total, enabling the creation of 270 unique sub-archetype combinations. Deployed in a lower-order 5R1C model, the archetypes effectively improve the “initial guesses” for building energy model inputs and enhance modeling accuracy and applicability. Potentially generalizable to all climate zones and building typologies in the US, the proposed workflow for UBEM archetype generation supports decision making in building retrofit planning and accelerates building stock decarbonization.