Stochastic activity-based approach of occupant-related energy consumption in residential buildings.

Summary The building sector is considered a major energy consumer and source of pollution among all economic sectors. It accounts for between 16 and 50 percent of national energy consumption. Reducing energy consumption and emissions is therefore an important step towards sustainable development. Recently, the transition to low-energy building construction has led to new performance and sustainability requirements, and thus further complicated the building design process. The behavior of the occupants is now considered as a determining factor of the energy performance of a building, especially in the case of low energy buildings (BBC). However, the simulation tools used in the building industry are not currently able to provide reliable estimates of occupant energy demand. Therefore, energy and building experts are paying great attention to developing more accurate methods for modeling and predicting the influence of occupants on building performance. These models must be able to provide more accurate estimates of energy consumption and assess the variability of these consumptions. Consequently, the objective is to allow building experts to improve their technical solutions, to improve the performance of their services, and to promote better targeted incentives to users in order to reduce their energy consumption. The objective of this thesis is to propose a model to estimate the energy consumption related to the behaviors of the occupants of residential buildings, taking into account the variability of consumption patterns through the diversity of socio-demographic and economic profiles of the occupants. A stochastic approach based on the notion of activity is therefore adopted. With this model, the energy consumption of a household is estimated by adding the energy consumption of the different domestic activities (such as cooking, washing clothes, etc.). The stochastic nature of the model is due to the probabilistic relationships established between household attributes on the one hand (household type, number of occupants, etc.) and the ownership of household equipment, appliance characteristics, their power, and activity quantities on the other hand. In order to establish these stochastic relationships, a sufficient number of attributes are taken into account to characterize a household. The proposed model has been applied for two household activities, namely watching television and washing clothes. Monte Carlo simulations are performed to provide energy consumption estimates for these two activities in three cases: for a specific household, for randomly generated households with constraints on their attributes, and for completely random households representative of the French population. A comparison between the results of the model simulation on the one hand and real energy consumption data on the other hand, allowed to validate the model for the two activities considered. A framework for the generalization of the model for other domestic activities was introduced, and its possible integration in the building design process was discussed and illustrated through a number of examples.