WINDOW OPENING BEHAVIOUR IN A NATURALLY VENTILATED SCHOOL

Abstract

Several studies have highlighted the significant disparities between the predicted performance of new buildings and their post occupancy measured performance. Occupant window opening behaviour has been shown to have a significant impact on both classroom indoor air quality and building energy use; and is currently poorly represented in building simulation tools. A post occupancy assessment of a new primary school was performed over a period of over one year. Concurrent measurement of window open state, CO2 concentration, temperature, and exterior environmental conditions were taken at a frequency of two minutes. In addition, classroom daily occupancy levels and monthly building energy usage were recorded. A probabilistic model of the proportion of windows open throughout the day as the occupants interact with the windows was developed based on the results of multinomial logistic regression analysis. The model was used to schedule window opening in the EnergyPlus simulation program. Predictions of both CO2 concentration and building energy performance, using the occupant behaviour model, were shown to give more accurate predictions than a model based on temperature set points.