Abstract
The simulation codes are diffusing widely in building design and energy performance evaluation. They also allow to refine the steady-state methods, particularly the estimations of the utilization factors, according to the technical standard EN ISO 13790:2008. The necessary validation steps used for diagnostic purposes are not aimed to evaluate the agreement of the results over the spectrum of configurations and conditions in real applications. Moreover, the relative impact of a single parameter on the global performance can be investigated only with an appropriate choice of the set of configurations analyzed. In this work, a method to evaluate the relative inaccuracy arising from the choice of a specific simulation code has been applied to compare two well-known dynamic simulation software applications: TRNSYS 16.1 and EnergyPlus 7. The main envelope variables investigated are dispersing surface amount, wall insulation and heat capacity, glazings insulation and solar transmittance, glazing dimensions and orientation, and internal gains. More than 1600 configurations were obtained through a full factorial plan for two Italian climatic conditions. Monthly heating and cooling needs, heating and cooling peak loads, and peaks occurrence time have been considered. Inferential statistics allowed weighting of the different variables impact on the found deviations.