The effect of weather datasets on building energy simulation outputs

Results of dynamic energy simulations of buildings are affected by many uncertainties, which are the main reason of the performance gap registered between simulated and operational performance. They depend mostly on the incorrect modelling of building components and their properties, the inadequate characterization of operational schedules, the limitations in the simulation algorithms used by energy simulation software, the quality and reliability of data contained in weather files. The first three limiting factors are somehow under the control and capacity of the person in charge of the simulation, that, nevertheless, may not always be able to get detailed building specifications to identify the correct set-points and schedules, or to choose an alternative simulation software. The information contained in weather datasets are, however, completely out of the control of the person in charge of the simulation that may only assume them as a boundary condition. Unfortunately, not all the weather databases show the same level of data accuracy; moreover, they may refer to a climate that substantially changed in the last decades. The effects on building energy simulation results, played by different weather files referred to the city of Milan, is showed and discussed, highlighting the substantial performance difference depending on them.