Assessment of thermal comfort at the building level: Evaluation of aggregation methods with a Danish case study of a campus building

Six methods for aggregating local thermal comfort scores in six offices to a single global score are investigated. using data collected by the building management system in a campus building of Aalborg University. Three static: 1) number of rooms weighted mean, 2) area-weighted mean, 3) desk-weighted mean, and three dynamic: 4) simple occupancy-weighted mean (PIR sensors), 5) advanced occupancy-weighted mean (PIR sensors mixed with the number of desks), 6) number of the occupants-weighted mean (camera readings). A notable disparity emerged between static methods, which rely solely on fixed parameters, and dynamic methods, which account for time-dependent factors over short timeframes. Dynamic methods consistently yielded lower global scores, irrespective of individual room performance. The difference can be up to 15% monthly. The PIR sensors, which are now commonly used in office or education buildings to control artificial lighting are a good indication of the occupancy (only present and not present). The information on number of occupants in the offices, collected by installed cameras, did not provide significantly better results in the analysed case study.