Nowadays, simulation tools are widely used to design buildings since their energy performance is increasing. Simulation is used to predict building energy performance and to improve thermal comfort of occupants, but also to reduce the environmental impact of the building over its whole life cycle and the cost of construction and operation. Simulation becomes an essential decision support tool, but its reliability should not be ignored. Hypothesis, made 10 years ago for buildings conception, are often not adapted to the new constructions because of physical phenomena which until now were overlooked. At the same time, guarantees of energy efficiency, which aims to check if actual energy performances are matching the conception goals, are becoming important. But there are usually differences between measured and simulation data. They may be the result of mistakes and unknowns on input parameters, on schedule occupation or on weather data. Today it's important to evaluate simulation and measurement reliability and uncertainties to improve design building. This PhD work aimed to evaluate and order simulation results uncertainties during the design building process. A methodology in three steps was developed to determine influential parameters on building energy performance and to identify the influence of these parameters uncertainty on the building performance. The first step uses the local sensitivity analysis and identifies the most influential parameters on the outputs among all parameters. This step enables to reduce the number of parameters which is necessary to proceed the following steps The second step is an uncertainty analysis focuses on quantifying uncertainty in model outputs. This step is conducted with the Monte Carlo probabilistic approach. The last step uses global sensitivity analysis which is the study of how uncertainty in the output of a model can be apportioned to different sources of uncertainty in the model input. This methodology was applied to the INCAS experimental platform of the French National Institute of Solar Energy (INES) in Le-Bourget-du-Lac to identify measure uncertainties and uncertainties on simulation hypothesis. This methodology may be used during the whole building design process, from the first sketches to the operating phase. It will enable to guide the architectural and technical choices and to avoid unstable options with important uncertainty. During the exploitation stage, this methodology will allow to identify the most suitable measurement in order to reduce parameters uncertainties and consequently to get the energy diagnostic more reliable. Moreover this methodology could also be used to determine uncertainties on related to inoccupants and to weather conditions.
