Transferring energy signatures across space and time to assess their viability for rapid urban energy demand estimation

Abstract

Energy signatures offer a rapid bottom-up energy modelling tool that relates thermal building space heating and cooling demand to ambient temperature. While simplified from a building physics perspective, signatures can easily exploit existing buildings’ operational datasets for their calibration, fitting individual buildings or archetypes. To assess their viability in modelling buildings’ heating and cooling demand at urban or larger scales, it is important to assess whether the same building model expressed as an energy signature could be used under different weather conditions, through spatial (location) or temporal (future weather) transfer. As energy demand depends on environmental factors other than temperature, a systematic error quantification from transferring energy signatures across different climatic contexts is required. These errors are assessed by comparing signature-based estimates to EnergyPlus simulations of a single-family home and an office building archetype exposed to current and future climate across eight US climate zones. Energy signatures are confirmed to be a valid approach if sufficient heating and cooling data is used for their identification. On an annual basis, signature-based estimates result in an error of approximately ± 30 % if the signature is spatially transferred, which can exceed 75 % for unsuitable source locations. When energy signatures are temporally transferred, an annual error of only ± 5–10 % is expected. At a daily resolution, the inherent error from applying the energy signature method is significantly larger than the additional error introduced through spatiotemporal transfer of the energy signature.