Investigating the probability of behavioural responses to cold thermal discomfort

• Empirical study design to monitor behavioural responses to thermal discomfort.
• Statistical tools to investigate adaptive behaviours to cold thermal discomfort.
• As internal air temperature decreases variations and levels in activity increase.
• Delay in clothing thermal adaptation.

In buildings, occupant behaviour is recognised as a major contributing factor to energy demand and in particular to heating consumption. To achieve thermal comfort within the heating season, people report to use heat in very different ways; for example behaviours include switching on the heating system, putting on warm clothes, drawing curtains, changing rooms, making a hot drink and using a hot water bottle. While research has focused on subjective accounts using interviews, diaries and questionnaires, little is known about the frequency and probability of these behaviours. Using a mixed-method approach, this paper reports on the results of a field study in dwellings using wearable and environmental sensors. The analysis investigates the probability of these behavioural responses as a function of seven independent variables; (1) external and (2) internal monitored temperature, (3) probability of heating being on or off, (4) time of the week, (5) time of the day, (6) the three categories of the predictive thermal comfort model, and (7) the three categories of the adaptive thermal comfort model. Results show that participants were more likely to increase their clothing and activity level as internal temperature decreased, although there was no significant change in activity level throughout the course of a day. Methodologically, this paper demonstrates the effectiveness of different statistical tools in analysing occupants’ behaviours. Substantively, this paper emphasises the need for future research to gather objective data on what people do.