Civil Engineering Reference
In-Depth Information
simulation software and is widely used by simulation experts and beginners alike.
The reason for this is that EnergyPlus is based on the most popular features and
capabilities of BLAST and DOE-2 and because it is only a calculation engine
without any graphical user interface. An example is the DesignBuilder (Tindale
2010
),
which
is
acknowledged
as
the
most
comprehensive
interface
for
EnergyPlus.
Building performance simulation has become an accepted method of assess-
ment during the building design process (Hoes et al.
2009
). Furthermore, due to an
increasing complexity of building design and higher demand for performance
requirements on sustainability, use of building simulations will become inevitable.
Mahdavi suggests that user interactions are difficult to predict at the level of the
individual (Mahdavi and Pröglhöf
2009
) although, behavioural patterns for
building occupants could be extracted from long-term observational data. Since
this data are the outcome of actual observations in a building, they are more
reliable than the currently applied simulation assumptions (Mahdavi and Pröglhöf
2009
). In this way, patterns obtained for one building cannot be transported to
other different buildings, because the occupancy models would be different.
Therefore, it is assumed that user behaviour is one of the most important
parameters influencing the results of building performance simulations. Unreliable
assumptions regarding user behaviour may have large implications for such
assessments. Moreover, this effect will become crucial when the design under
investigation contains improved passive and active energy efficiency measures.
In current building performance simulation tools, user behaviour is mirrored in
a very rigid way. In recent years, some models have been developed to include the
interaction of the user behaviour in building simulation. Models for the simulation
of occupant interactions with windows have been addressed in the Humphreys
algorithm for window opening that was derived from analysis of extensive survey
data and was implemented in the ESP-r software (Rijal et al.
2007
). More work
related to window usage can be found in Herkel et al. (
2008
), Frederic and Darren
(
2009
). In (Borgeson and Brager
2008
), a methodology for predicting occupant
window control is presented. Reinhart developed LIGHTSWITCH-2002 using a
dynamic stochastic algorithm (Reinhart
2004
). Based on an occupancy model and
a dynamic daylight simulation application, it predicted manual lighting and blind
control actions providing the basis for the calculation of annual energy demand for
electrical lighting. Page hypothesised that the probability of occupancy at a given
time step depends only on the states of occupancy at the previous step (Page et al.
2008
). In this way, he proposed the application of Markov chains towards occu-
pancy prediction.
As a continuation of the work by Reinhart and Nicol (Nicol
2001
), Bourgeois
attempted to bridge the gap between energy simulation and empirically based
information on occupant behaviour via a module called SHOCC that was also
integrated in ESP-r application (Bourgeois et al.
2006
). The SHOCC module is
applied in Hoes et al. (
2009
); however, this module requires several assumptions
with respect to the degree of occupation and the behaviour of users. That is, when
the number of occupants is relatively high and they operate with less strict
