Statistical analysis of the ranking capability of long term thermal discomfort indices and analysis of sensitivity to time related boundary conditions

The scientific literature and some standards offer a number of long-term thermal discomfort indices (LTDIs) and methods for predicting the likelihood of summer overheating in buildings. Such metrics can be useful tools for driving the optimization process of the design of new buildings and for the operational assessment of thermal comfort in existing buildings. Focusing on the summer period, in the first part of this thesis 16 LTDIs are applied for assessing a sample of different variants of a large office building. Such building variants are obtained by varying in discrete steps some key design parameters of the building envelope. The values of the 16 indices are compared and contrasted in subsets to identify similarities and differences in assessing the whole sample of building variants. The indices deliver significantly different results with respect to the same building variant and identify diverse optimal building variants. Accordingly, the choice of the long-term discomfort index has a strong impact on the outcome, and, therefore, this thesis is intended to provide clarification on how to employ them in a reliable and conscious manner. Besides, some of the analyzed indices have shown the capability to deliver a similar ranking even based on different comfort models. Furthermore, LTDIs depend on a number of boundary parameters that are currently not univocally defined in standards, thus introducing an uncertainty when they are calculated. The second part of this thesis is focused on studying the effect on the outcome of the arbitrary choice of one of such boundary conditions, i.e. the calculation period. This condition has been translated into two sources of variation: the time interval defined as summer and the daily hours of occupation of the building. Then a selection of 6 long-term discomfort indices has been tested over a number of building variants and over all the combination of the defined cases of the two considered sources of variation. As a result it was found that this incomplete definition of the calculation period may lead to different classification of the same building. There appears to be a need in Standards for an explicit definition of the calculation period.