Type:

Year:

Supervisor:

Institution:

Page(s):

Abstract:

The building sector is considered as a major energy consumer and pollution source among all economic sectors. It accounts for important shares, ranging between 16 and 50 percent, of national energy consumption worldwide. Reducing these consumptions and emissions is thus an important step towards sustainable development. Recently, the shift towards constructing low-consuming and nearly zero-energy buildings lead to further requirements with regard to performance and sustainability, and thus caused the design process of buildings to be more complex. Occupants’ behavior is now considered as a key determinant of building’s energy performance especially in the case of green buildings. Yet, energy simulation tools used in buildings industry nowadays are not capable of providing accurate estimations of occupant-related energy demands. Therefore, buildings and energy experts are devoting considerable efforts on developing more precise methods for modeling and forecasting occupants influence on whole building performance. Such models can provide accurate energy estimates and can assess future consumption variability. Consequently, building experts may improve their technical solutions, ameliorate their service performances, and promote targeted incentives.
The objective of this dissertation is to propose a model for forecasting occupant-related energy consumption in residential buildings, while accounting for variability in consumption patterns due to diversity in occupants’ socio-demographic and economic profiles. A stochastic activity-based approach is thus adopted. By activity-based, it means that energy consumption of a household is estimated by summing up the energy use of different activities performed (such as cooking, washing clothes, etc.). The stochastic nature of the model is due to the probabilistic mapping established between household attributes from one side (household type, number of occupants, etc.) and the corresponding appliance ownership, appliance characteristics and power rating, and activity quantities from the other side. In order to establish these stochastic relations, a fairly sufficient number of households’ characterizing attributes is taken into account. The proposed model is applied for two domestic activities, namely watching TV and washing laundry. Three types of Monte Carlo simulations are performed to provide energy estimates for these two activities: for a given specified household, for randomly generated households with constraints, and for totally random population-wise households. A comparison between model’s simulation results and real measured energy consumption data enables validating the model for the two considered activities. A generalization framework of the modeling approach for other domestic activities is sketched, and its possible integration into buildings design process is discussed and illustrated through a number of examples.

Keywords:

Energy consumption, residential building, energy model, green building, energy performance, building occupants, household profile, domestic activity, domestic appliance, occupant behavior, consumption variability, energy simulation