PNNL

Abstract

Building energy simulation play a significant role in building design and retrofit. Most applications deal with one or few individual buildings which enables the specification of detailed model inputs. However, building energy simulation can also be a powerful tool for assessing energy performance even when detailed building characteristics are not available. In this study, field data were collected on a limited randomly selected new residential homes in 8 US states aiming to evaluate energy code compliance and saving potential. The limited data prohibits deriving model inputs for the homes visited let alone for the entire unknown new home stock. We use prototype building to construct a large number of building models and utilize bootstrap sampling to draw model inputs from the limited data. We demonstrate that overall energy performance of a large population of new homes can be assessed by the novel framework through EnergyPlus simulation using limited field data.