Building model calibration using energy and environmental data

A large number of randomly interacting variables combine to dictate the energy performance of a building. Building energy simulation models attempt to capture these perturbations as accurately as possible. The prediction accuracy of building energy models can now be better examined given the widespread availability of environmental and energy monitoring equipment and reduced data storage costs. In this paper a set of two calibrated environmental sensors together with a weather station are deployed in a 5-storey office building to examine the accuracy of an EnergyPlus virtual building model. Using American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) Guide 14 indices the model was calibrated to achieve Mean Bias Error (MBE) values within ±5% and Cumulative Variation of Root Mean Square Error (CV(RMSE)) values below 10%. The calibrated EnergyPlus model was able to predict annual hourly space air temperatures with an accuracy of ±1.5 °C for 99.5% and an accuracy of ±1 °C for 93.2% of the time.