The aim of this thesis is to achieve better knowledge
of how occupants and weather affect the
predictions from building performance simulations.
The influence of weather is evaluated
based on robustness and design tendencies, while
the occupancy modelling is evaluated on building
performance and dimensional values. A series
of weather files and methods to perform
occupancy modelling are tested with factorial
experiment. These are performed on either
building performance simulation tool, Be15 and
BSim, respectively.
Based on a fictional office building, stochastic
simulations were carried out with two simulation
methods. The whole building is simulated
in Be15, where the influence from including multiple
historic weather files are compared with
results from Design Reference Year. The design
solutions were evaluated in terms of their
robustness level towards weather files and the
most important input parameters were identified
through sensitivity analyses. Additionally the
design tendencies that grants a desired building
performance are studied.
Furthermore, for a selected room modelled in
BSim, the influence of using stochastically generated
occupancy profiles were studied and compared
to static occupancy profiles in terms of
performance and system dimensioning. Tendencies
that leads to deviations in the building performance
and system dimensioning are outlined.
Sensitivity analysis were conducted to identify
the influence from variability of occupants and
weather on the evaluated parameters.