Massimo Fiorentini

Fixed thermostat setpoints and schedules are commonly used in residential building energy simulation and rating. While this approach is simple to implement, it does not represent occupants with varying preferences. In this study, based on field data from 102 households in three Australian cities, two alternative thermostat setting approaches were investigated. The first method (Probability Distribution Approach) uses all the values in a thermostat settings probability distribution generated from the field data. This was compared with a more straightforward method, where the thermostat settings were derived by applying weighted average thermostat settings. Both approaches were benchmarked against a series of simulations that used randomly generated thermostat settings with the same thermostat settings probability distributions. Results show that the Probability Distribution Approach matches better the benchmarking results (CV(RMSE) 1-8%) than the weighted average method (CV(RMSE) 9-37%), particularly for cooling demand.