The literature states that the occupancy and related operational
characteristics in buildings are key variables that cause the gap between the estimate
and actual thermal and energy performance. To address such issue, the objective of this
study is to investigate the uncertainties of occupant behaviour in building performance
simulation through a probabilistic approach. This case study considers a model of a
low-income dwelling in southern Brazil using five different construction for the
envelope and with natural or hybrid ventilation. Field survey provided a dataset of
uncertainties of the occupant behaviour, which was related to the occupancy of the
rooms, operation of openings and use of electric appliances. The EnergyPlus
programme was used to conduct the simulations and the R Studio was used for data
processing, analysis, and treatment. A global sensitivity analysis was performed, along
with an uncertainty analysis. The results showed that the number of occupants, the
schedules of occupancy of the bedrooms, the setpoint temperatures for operating the
openings, the cooling setpoint of the HVAC and the limits for operative temperatures
of the rooms were the most influent variables for the thermal and energy performance,
especially in the heating period. The uncertainty was up to 65.6% for estimating the
degree-hours for heating (in the natural ventilation mode) and up to 59.3% for
estimating the total electricity consumption with HVAC (in the hybrid ventilation
mode), indicating that these operational uncertainties had a great impact in the
simulation results.
Keywords: Building simulation, EnergyPlus, Energy consumption, Global
sensitivity, Operational uncertainty, Performance evaluation, Sensitivity analysis,
Sobol’ indices, Thermal performance, User behaviour, Uncertainty analysis.
