Civil Engineering Reference
In-Depth Information
schedules, the predictions are less reliable. For these, and more complex situations,
the USSU module can be applied (Tabak and Vries
2010
).
The ECM proposed in this chapter is constructed with generated data, simu-
lating a building, and highlights from current simulation tools that rely on
assumptions or simulation methods. The applicability of the results is limited to
supporting or accurately predicting the effectiveness of energy efficiency measures
in a building. The purpose of the ECM is to fill that gap. Furthermore, the ECM
can be used to predict the outcome of the installation of energy efficient tech-
nologies by exploring alternative scenarios.
2.2 User Behaviour Modelling
The study of human behaviour has been the focus of many fields of research and
has been extensively investigated in business process modelling, cognitive mod-
elling, distributed artificial intelligence, computational organisational theory and
educational psychology (Atallah and Yang
2009
). The main challenge is related to
the modelling of complex human behaviour using realistic yet adaptable models.
One of the most popular models is that of a hidden Markov model (HMM)
(Rabiner
1989
), which is a stochastic model. Due to their ability to model spatial-
temporal information in a natural way, a significant amount of work in the area of
behaviour recognition is based on HMMs.
The concept of energy behaviours is explored in Lopes et al. (
2012
) where the
interactions between energy behaviours and energy efficiency in buildings are re-
viewed from an interdisciplinary to multidisciplinary end-use energy efficiency in
buildings and modelling energy behaviour trends. A survey in behaviour model-
ling is performed by Atallah and Yang (
2009
). Because little is known about user
activity in buildings, several experiments observe motions of real users with
cameras and other means of locating people (Tabak
2009
). In other experiments,
user activities regarding light control have been monitored providing useful sta-
tistics to improve lighting control (Mahdavi and Pröglhöf
2009
). Although such
experiments provide realistic user patterns, it is not easy to extract the reasons for
activities and thus derive abstract dynamic models. The study performed in
Borgeson and Brager (
2008
) classifies human behaviour as not deterministic. In
(Zimmerman
2007
), agent models are addressed as suitable for behaviour mod-
elling, and multiple-agents simulation studies are performed. Other agent-based
model,
which
takes
into
account,
the
building
occupant's
decision-making
regarding energy usage is performed in Chen et al. (
2012
).
Another important modelling technique is applying pattern recognition analysis
(Abreu et al.
2012
) in order to identify common behaviours, which could provide
valuable feedback about dominant behaviours (affect building energy consumption
directly) and thus improving the effectiveness of energy-efficient technologies. In
order to isolate the effects of occupant behaviour on building energy consumption
