Cluster Analysis of Masonry Types to Analyze the Energy Performance of Historic Buildings: Preliminary Results

Abstract

The study discussed in this paper is part of a research, which applies the method of masonry typology to the historic building stock of Sicily. The re- search is being carried out on a set of local contexts, currently 14 villages and small towns. The general aim is building a database of mechanical and thermal data specific to historic masonry, and verifying if correction factors, based on the construction features of local masonry types, may improve the reliability of calculated U-value. The cluster analysis described in this paper was performed to examine the technical data collected for each type, in relation to the thermal modeling of historic stonewalls. In this cluster analysis, each masonry type was treated as an observation, whose quantitative and qualitative variables are the construction details of the local masonry types. A non-hierarchical method was applied through the functions Daisy and PAM in software R. The clustering structure is made of two clusters and, though featured by the rate of reliability SC = 0.33, it provides relevant indications about the construction characteristics to select as parameters in the thermal models of local masonry types. The paper discusses these preliminary findings and their further development.