Existing dynamic energy simulation tools, developed in thè last decade, exceed thè static size of thè simplified methods through a better and more accurate prediction of energy use; however, they are stili unable to replicate the actual dynamics that govern energy uses within buildings. Energy performance is dramatically affected by thè behavior of users inside buildings: each architectural project is proposed on thè base of a singular spatial organization and occupier’s usability, but once thè building-plant system has been realized, this could be used differently from the assumptions made in the design phase. Moreover, the use of imprecise and standardized occupant behavior can blur the effect of other energy saving solutions.

The inappropriateness of deterministic assumptions about occupants’ interactions with building's Controls is proved by the field evidence: despite on-going improvements in the building envelope thermal performances, energy demand has not decreased in the expected way. Moreover, recent studies compare deterministic simulation results to the simulation results where stochastic models are implemented, showing a significant energy use increase when switching from one approach to another [(Fabi et ai, 2012), (D’Oca, 2012), (Buso, 2012)]. For these reasons, a stochastic simulation of the designed edifice, in which statistical methods can be used to generate individual occupancy patterns, indeed of fixed schedules, is needed.

Several investigations have been conducted inliterature in order to comprehend thè most influencing drivers related to occupant behavior in an office building. Accordingly to previous researches (Annex 53, ECBCS), is possible to highlight that users principally restare their comfort condition by opening and closing windows. It is therefore important to take window opening behavior into account, when designing energy efficient buildings.

Given this background, the main purpose of thè presented research is to investigate the implications of different assumptions with regard to window operation in a mechanically and naturally ventilated office building for thè energy use and located in different locations, i.e. Turin and Athens, with the aim to evaluate the influence of climate on the predicted energy consumption. Toward this end, a dynamic numeric simulation application, which necessarily takes into account thè possibility that an action is carried out through statistical input parameters, was deployed to simulate the building thermal performance. This new method defines behavioral models by using an algorithm based on a statistical data set built on monitoring occupants’ real interaction within the building-plant system, published by Haldi and Robinson (2009). Then, an individual behavioral model approach, taken from Parys’ (2013) PhD thesis, was developed by implementing the logistic regression models for ac live and passive users with the purpose to compare an aggregated behavioral model to a single one, where the differences between users are displayed.

The method of the research developed in this master thesis is based on the assumption that only switching from an automated approach to a manual one, it will be feasible to obtain energy consumption prediction closer to reality, already in the design phase.