Statistical and Stochastic Modelling of French Households and Their Energy Consuming Activities

Guillaume Ansanay-Alex (1) , Yassine Abdelouadoud (1) , Pascal Schetelat (1)

1 CSTB - Centre Scientifique et Technique du Bâtiment

There often is a significant difference between predicted energy performance of buildings and measured energy performance once buildings are operational. Actual occupant behavior is often different from the assumptions made in the design stage, and this is regarded as one of the main reasons for the performance gap. The diversity in occupants' sociological and behavioral characteristics must then be taken into account when predictions are undertaken, in order to determine confidence intervals of the expectable performances. Software tools must then be made available to design firms, letting them generate statistically representative sets of occupants and plausible activity and appliance use scenarios. Here we introduce two tools developed at CSTB, named Qiriel and Croniq, focused respectively on statistical generation or inference of dwellings, appliances and households, and on stochastic simulation of occupants' activities and the implied appliances activation and power consumption. They are built upon French national databases on households and time use surveys. These tools enable their users to generate in a short amount of time hundreds of occupancy, activity and power consumption profiles for specific sociological profiles or for representative sets of households. The obtained sets of series can then be used to determine median and extreme configurations of building usage, and the implied range of building performances. Such tools can also be useful at the building design stage, in order to evaluate how robust a building, its HVAC systems and their regulation can be when confronted to a large variety of usage patterns.

Mots clés: energy performance gap; occupant behavior; prediction; building design

Domaines: Eco-conception; Applications [stat.AP]

Guillaume Ansanay-Alex, Yassine Abdelouadoud, Pascal Schetelat. Statistical and Stochastic Modelling of French Households and Their Energy Consuming Activities. 12th REHVA World Congress - CLIMA 2016, May 2016, Aalborg, Denmark. ⟨hal-01308929⟩