In Europe, households consume about 27% of the final energy consumption. To reduce the energy household consumption, energy performance regulations set minimum requirements for the energy efficiency of homes. The calculation method for the energy performance focusses mainly on building characteristics such as the level of insulation and the energy efficiency of the ventilation system. However, buildings don't use energy, people do. The way households behave in their homes has an significant impact on the energy consumption as well. Currently, this user behaviour is hardly taken into account in the calculation method.

This dissertation focusses on one aspect of user behaviour, namely electricity consumption. To understand household electricity consumption, we need to know where people are, what activities they are doing and which appliances they are using to do these activities. To this end, the dissertation presents three probabilistic models which forecast the presence of people, their activities and their electricity use. The models also take into account household characteristics that have an important impact, such as the number of household members, their age and their employment.

The models may be useful for building designers, policy makers and households. Building designers need better predictions of the energy performance of the buildings, especially when designing nearly-zero-energy homes. Policy makers need insight in the behavioural aspects of energy consumption when outlining their decisions regarding the allocation of subsidies or the tightening of building performance regulations. Households can save on their energy bills by receiving better feedback which compares them to similar households or provides appliance-specific information.

research funded by Prospective Research for Brussels (Innoviris)

Download PhD dissertation