“Mind reading” building operation behaviour

Modern buildings are operationally classified as fully naturally ventilated (free running, FR), fully mechanically ventilated and conditioned (air-conditioned, AC) or intermediary (mixed mode, MM). Each strategy implies (i) the selection of different internal comfort models and controls, e.g. adaptive thermal comfort + operable windows in FR and static thermal comfort (Predicted Mean Vote - PMV) + non-operable windows in AC, and (ii) different energy consumption and carbon emissions with FR being the least (zero) and AC being the most. In each instance, the choice of strategy is dictated by the designer with little choice for the occupants. Here we ask the question: what operational mode would occupants select, given complete freedom? We examine this question using real concurrent data from three offices of similar size and use. In each, occupants have complete freedom in adopting FR, AC or a series of intervening MM strategies. We use monitored data on window and air-conditioning operation combined with internal and external thermal conditions to create validated computer models for each office. We discover the true operational mode of each office by comparing the computer model against 23 different scenarios using Dynamic Time Warping for binary (window open/close, AC on/off) and Euclidean distance for continuous operative temperature time series. Strikingly, results demonstrate that while each office used divergent strategies across the seasons, the indoor conditions were very similar and attainable through NV alone. This suggests that while a purely NV strategy is likely to deliver indoor thermal comfort, understanding occupant motivation and educating them on the impact of AC operation is needed to minimise energy use.