Investigating thermostat setpoint preferences in Canadian households

Occupants' thermostat setpoint preferences play a vital role in HVAC systems' operation and significantly influence the building energy performance. However, despite the diversity in indoor temperature preferences, most building energy codes assume identical thermostat setpoints for buildings of the same type. To this end, this study aims to demonstrate the variations in temperature setpoint preferences across Canadian households by analysing thermostat data collected from ~13,000 residential buildings. The objectives of this study are to (1) determine the average heating, and cooling thermostat setpoints in residential buildings, (2) rank the importance of different attributes that influence setpoint preferences, and (3) extract distinct heating and cooling setpoint profiles. Statistical methods were used to identify the average thermostat setpoints in different provinces. A random forest ensemble learning model was then used to rank the relative importance of different attributes on the setpoint temperatures. Finally, the k-Shape clustering technique was used to extract distinct heating and cooling setpoint temperature profiles. The obtained results were compared with the building energy codes, standards and differences up to ~3°C were found relative to code assumptions.