Exploring occupants' impact at different spatial scales

Buildings' users have widely been accepted as a source of uncertainty in building energy performance predictions. However, it is not evident that the diversity of occupants' presence and behavior at the building level is as important as at the room level. The questions are: How should occupants be modeled at different spatial scales? At the various scales of interest, how much difference does it make if: (1) industry standard assumptions or a dynamic occupant modeling approach is used in a simulation-based analysis, and (2) probabilistic or deterministic models are used for the dynamic modeling of occupants? This paper explores the reliability of building energy predictions and the ability to quantify uncertainty associated with occupant modeling at different scales. To this end, the impacts of occupancy and occupants' use of lighting and window shades on the predicted building lighting energy performance at the room and building level are studied. The simulation results showed that the inter-occupant variation at larger scales is not as important as at the room level. At larger scales (about 100 offices), the rule-base model, custom schedule model, and stochastic lighting use model compared closely for predicting mean annual lighting energy use.