A Framework for Meta-Analysis of the Role of Occupancy Variables in the Energy Use of Commercial Buildings

Occupan ts interact with buildings in various ways via their presence (passive effects) and control actions (active effects). Passive effects regard people’s presence as a heat (sensible and latent) generation source while active effects take into account occupants’ operation of appliances and control of lighting, heating, cooling, and ventilation devices. Therefore, understanding the influence of occupants is essential if we are to evaluate the performance of a building. Conventional methods represent complex relationships between buildings and occupants by simplified diversity profiles, which are inadequate in cases where an accurate assessment of the building energy consumption is required. Current state-of-theart methods aim at reproducing the stochastic nature of occupant presence and behavior. They can become overly detailed and are hard to generalize to future non-standard buildings when there are no data to rely on for the generation of the stochastic model. In this paper, we propose a time series model for the variability of the two major occupancy variables: presence and actions, and develop the framework for a meta-analysis that synthesizes occupancy data gathered from a pool of buildings. We then discuss the appropriate level of granularity for occupancy modeling with respect to various outcomes of interest such as energy consumption and electricity peak demand via a sensitivity analysis. Our results show that time series models are able to generate realistic variability of occupancy variables, requiring only input that specifies standard diversity profiles from the energy modeler. Along with the proposed metaanalysis, we will be able to generalize previous research results and enable statistical inferences to choose occupancy variables for future buildings. The sensitivity analysis shows that the variability of occupancy variables does not significantly affect predicted energy consumption on the premise that there is sufficient knowledge about the typical presence and usage of the building. It is a different story when it comes to the electricity peak demand, where the stochastic nature of occupancy variables certainly matters.