Including measurement uncertainty in building energy performance calculation methods

In building design, energy performances are commonly predicted based on deterministic stationary or dynamic calculations. However, many contributing parameters are inherently uncertain, resulting in potentially unreliable values for the performance indicators. To overcome this, a probabilistic design method is recommended to take uncertainties into account. Such uncertainty-based optimisation often requires many simulations, making it extremely time-consuming. To avoid this, meta-modelling can be of interest. A meta-model mimics the original numerical model with a simplified fast model.  The current paper presents a global probabilistic design methodology to take different kinds of uncertainty into account in building energy performance calculation methods.