On the potential of clustering approaches for hygrothermal material properties based on three degradation risks in solid masonry constructions

Abstract

Historic masonry has a rich and colorful history making it a treasured part in our society. To preserve and protect this heritage, adequate moisture control, retrofit, and restauration strategies are required. However, due to the large range of material properties inherent to historic brickwork, a single uniform renovation strategy appears impossible. To describe similarity in brickwork, the existing clustering approach developed by Zhao was evaluated. The idea is that different types of bricks with similar properties can be represented by a single representative brick for that cluster, for example, when conducting hygrothermal simulations. It could help improve existing retrofit practice by reducing characterization processes and minimizing time-consuming laboratory measuring tests. However, in this paper the approach presented by Zhao is questioned since the clustering is solely based on an equal impact of the material properties and the response behavior and associated degradation risks are neglected. The aim of this paper was twofold. Firstly, similarity in brickwork obtained by clustering according to Zhao was evaluated by means of hygrothermal simulations to see whether bricks in the same cluster show similar degradation risks. Zhao’s clustering provides homogenous clusters regarding physical material properties, but significant variation was found in degradation risks for different bricks within the same cluster. Secondly, a methodology is presented to translate similarities in degradation profiles toward similarities in material properties. Sensitivity analyses were used to study the response behavior based on three degradations risks: mold growth, wood rot, and frost damage. Finally, an overall clustering scheme was generated for brickwork, based on classification trees for different degradation phenomena.