Building adaptation and re-use can contribute to a circular and sustainable built environment, as existing buildings are adapted and the need for new construction materials is reduced. The “adaptability” of buildings has been widely studied; however, few of these studies are quantitative. This paper uses Artificial Neural Networks (ANN) and Logistic Regression (LR) models to explore relationships between the physical features of buildings and their demolition or adaptation outcomes. Source data were taken from 59 buildings that were either demolished or adapted in the Netherlands. After the models were created and validated, a series of sensitivity studies were conducted to evaluate relationships between physical parameters and building outcomes. The physical parameter with the strongest relationship to adaptation outcomes was demountability (ease of removal) of building service elements. The quantitative results were then compared to results from an adjacent qualitative study. The relationships observed from the quantitative sensitivity studies align well with the qualitative observations.

No abstract