A case for thermostat user models

Today’s thermostat setpoint models naïvely assume fixed schedules, ignoring the reality of user control and its large variability. Better models must include more realistic user-behavior profiles to correctly evaluate the energy benefits of the next generation of thermostats against a realistic baseline. Data from a recent thermostat field study were analyzed to demonstrate the variation and patterns associated with manual adjustment of programmable thermostats and its consequences on observed and simulated energy consumption. A practical modeling technique for describing variable setpoint schedules was applied and compared with standard, fixed setpoint assumptions. Room air temperature data from 63 apartments near Boston, MA were used to generate unique hourly heating setpoint schedules. These observed temperature histories were then used to model the expected variation in energy use of an apartment due to manual thermostat adjustment. Significant differences in energy consumption were observed when variable setpoints were used instead of fixed setpoints, indicating the need for improving thermostat assumptions and updating models with more realistic schedules.