The use of cluster analysis to assess thermal comfort in university classrooms

Nowadays, providing health, well-being, productivity and energy efficiency to users inside buildings is essential. Applying these aspects aligned with sustainability becomes necessary to reduce the use of heating, ventilation, and air-conditioning (HVAC) systems. These systems are currently used to provide better thermal conditions to the occupants, who spend around 80% of their time indoors. The actual thermal conditions can be affected by several factors, such as the climatic type of the region, orientation, size, building type, and energy levels, among others. To assess thermal conditions inside buildings, several thermal comfort models have been developed over the years. However, the Predicted Mean Vote (PMV) created by Fanger is still the most common model to assess thermal comfort indoors. In this context, this research aimed to analyze thermal comfort conditions in university classrooms in Southern Brazil. By collecting the environmental and personal variables of thermal comfort and the mean thermal sensation of students through measurements and questionnaires, a total of 519 responses were obtained during the Brazilian autumn. A statistical cluster analysis was performed to classify individuals according to their sensations. Differences between genders were verified and changing indoor temperatures lower in winter would therefore save HVAC energy without impacting occupant comfort.