Simulating occupants' impact on building energy performance at different spatial scales

Abstract

It is not evident that modeling the stochastic nature of occupants' presence and behavior at the building level is as important as at the room level. Given this line of inquiry, the question is: how should occupants be modeled at different spatial resolutions? How much difference does it make to annual building energy use predictions if: (1) industry assumptions or dynamic models are used, and (2) probabilistic or deterministic models are used? This paper explores the reliability of predicting lighting energy use and the ability to quantify uncertainty at different scales. To this end, the impacts of occupancy and lighting and window shade use behaviors on lighting use predictions at various scales are studied. For occupant modeling, stochastic and rule-based models and custom and standard schedules are used. The simulation results indicated that the whole-building lighting energy use predictions with the stochastic models approached a consistent value of 2.3 kWh/m2 with office buildings larger than 100 offices. It shows that the impact of individual occupants diminishes as the office building size increases. This research concludes that the required office building size to provide a good approximation of the population is highly dependent on the occupant modeling approaches. The results reveal that the deterministic models (i.e. rule-based models and custom schedules) can reasonably represent occupants' impact on building energy performance at larger scales.